Changeset 33813

Show
Ignore:
Timestamp:
18.12.2019 21:38:44 (5 weeks ago)
Author:
ak19
Message:

With the bugfix from yesterday and the inclusion of http(s)://mi.* type URLs in setting the Websites mongodb collection's urlContainsLangCodeInPath property, and updated/improved mongodb queries and their results I have now regenerated the latest geojson json data and maps.

Location:
other-projects/maori-lang-detection
Files:
4 added
6 removed
14 modified
3 moved

Legend:

Unmodified
Added
Removed
  • other-projects/maori-lang-detection/MoreReading/mongodb.txt

    r33807 r33813  
    461461117496 
    462462 
    463 # Find number of websites who have 1 or more pages in Maori (a positive numPagesInMRI) 
     463# Find number of websites that have 1 or more pages detected as being in Maori (a positive numPagesInMRI) 
    464464db.getCollection('Websites').find({numPagesInMRI: { $gt: 0}}).count() 
    465465361 
     
    488488# Number of sites with URLs containing /mi(/) 
    489489db.getCollection('Websites').find({urlContainsLangCodeInPath:true}).count() 
    490 153 
     490X 153 
     491# Number of sites with URLs containing /mi(/) OR http(s)://mi.* 
     492db.getCollection('Websites').find({urlContainsLangCodeInPath:true}).count() 
     493670 
    491494 
    492495# Number of websites that are outside NZ that contain /mi(/) in any of its sub-urls 
    493496db.getCollection('Websites').find({urlContainsLangCodeInPath:true, geoLocationCountryCode: {$ne : "NZ"} }).count() 
    494 147 
    495  
    496 # 5 sites with URLs containing /mi(/) that are in NZ 
     497X 147 
     498# Number of websites that are outside NZ that contain /mi(/) OR http(s)://mi.* in any of its sub-urls 
     499db.getCollection('Websites').find({urlContainsLangCodeInPath:true, geoLocationCountryCode: {$ne : "NZ"} }).count() 
     500656 
     501 
     502# 6 sites with URLs containing /mi(/) that are in NZ 
    497503db.getCollection('Websites').find({urlContainsLangCodeInPath:true, geoLocationCountryCode: "NZ"}).count() 
    498 6 
     504X 6 
     505# 14 sites with URLs containing /mi(/) OR http(s)://mi.* that are in NZ 
     50614 
    499507 
    500508 
     
    609617 
    610618WORKS: 
    611 // count of country codes for sites that have /mi(/) in path 
     619// count of country codes for sites that have /mi(/) or http(s)://mi.* in URL path 
    612620 
    613621db.Websites.aggregate([ 
     
    743751# These are the TENTATIVE NON-PRODUCT SITES 
    744752# Should be less than the point 4, but more than 1 to 3 
     753 
    745754db.getCollection('Websites').find({$and: [{numPagesContainingMRI: {$gt: 0}},{$or: [{urlContainsLangCodeInPath: false}, {$and: [{urlContainsLangCodeInPath: true}, {geoLocationCountryCode: "NZ"}]}]}]}).count() 
    746 859 
     755X 859 
     756 
     757Now with http(s)://mi.* also excluded, the above query returns a count of: 
     758389 
     759 
     760 
     761BUT THIS IS THE CORRECT VERSION OF THE QUERY: 
     762db.getCollection('Websites').find({$and: [{numPagesContainingMRI: {$gt: 0}},{$or: [{geoLocationCountryCode: "NZ"}, {urlContainsLangCodeInPath: false}]}]}).count() 
     763389 
     764 
    747765 
    748766# 6. Now do the counts by country code of the above, by pasting the query of point 5 as the $match clause (i.e. without the .count() suffix) 
     
    913931    { 
    914932        $match: { 
    915             $and: [{numPagesContainingMRI: {$gt: 0}}, {geoLocationCountryCode: {$ne: "NZ"}}, {domain: {$not: /.nz$/}}, {urlContainsLangCodeInPath: {$ne: true}}, ] 
     933            $and: [{numPagesContainingMRI: {$gt: 0}}, {geoLocationCountryCode: {$ne: "NZ"}}, {domain: {$not: /.nz$/}}, {urlContainsLangCodeInPath: {$ne: true}}] 
    916934        } 
    917935    }, 
     
    926944    { $sort : { count : -1} } 
    927945]); 
     946 
     947 
     948We can knock of another 54 non-NZ sites with our new urlContainsLangCodeInPathPrefix field: 
     949 
     950   db.getCollection('Websites').find({urlContainsLangCodeInPathPrefix: true, geoLocationCountryCode: {$ne: "NZ"}, domain: {$not: /.nz$/}}).count() 
     951   54 
     952 
     953 
     954SO, can repeat query with new field "urlContainsLangCodeInPathPrefix": 
     955Number of sites containing >= 1 MRI sentences that are not from NZ or of .nz TLD and which don't contain "/mi(/)" or "http(s)://mi." in URL path: 
     956   db.getCollection('Websites').find({$and: [ 
     957                     {numPagesContainingMRI: {$gt: 0}}, 
     958                     {geoLocationCountryCode: {$ne: "NZ"}}, 
     959                     {domain: {$not: /.nz$/}}, 
     960                     {urlContainsLangCodeInPathSuffix: {$ne: true}}, 
     961                     {urlContainsLangCodeInPathPrefix: {$ne: true}} 
     962                ]}).count() 
     963 
     964   651 
     965 
     966 
     967REDO THE COUNT BY COUNTRY QUERY FOR THIS: 
     968 
     969db.Websites.aggregate([ 
     970    { 
     971        $match: { 
     972            $and: [{numPagesContainingMRI: {$gt: 0}}, {geoLocationCountryCode: {$ne: "NZ"}}, {domain: {$not: /.nz$/}}, {urlContainsLangCodeInPathSuffix: {$ne: true}}, {urlContainsLangCodeInPathPrefix: {$ne: true}}] 
     973        } 
     974    }, 
     975    { $unwind: "$geoLocationCountryCode" }, 
     976    { 
     977        $group: { 
     978            _id: {$toLower: '$geoLocationCountryCode'}, 
     979            count: { $sum: 1 }, 
     980            domain: { $addToSet: '$domain' } 
     981        } 
     982    }, 
     983    { $sort : { count : -1} } 
     984]); 
     985 
     986 
     987AFTER BUGFIX FOR miInURLPath being set at the correct now: 
     988db.getCollection('Websites').find( 
     989{$and: [ 
     990    {numPagesContainingMRI: {$gt: 0}}, 
     991    {geoLocationCountryCode: {$ne: "NZ"}}, 
     992    {domain: {$not: /.nz$/}}, 
     993    {urlContainsLangCodeInPath: {$ne: true}} 
     994]}).count() 
     995 
     996220 
     997 
     998db.Websites.aggregate([ 
     999    { 
     1000        $match: { 
     1001            $and: [ 
     1002            {numPagesContainingMRI: {$gt: 0}}, 
     1003            {geoLocationCountryCode: {$ne: "NZ"}}, 
     1004            {domain: {$not: /.nz$/}}, 
     1005            {urlContainsLangCodeInPath: {$ne: true}} 
     1006            ] 
     1007        } 
     1008    }, 
     1009    { $unwind: "$geoLocationCountryCode" }, 
     1010    { 
     1011        $group: { 
     1012            _id: {$toLower: '$geoLocationCountryCode'}, 
     1013            count: { $sum: 1 }, 
     1014            domain: { $addToSet: '$domain' } 
     1015        } 
     1016    }, 
     1017    { $sort : { count : -1} } 
     1018]); 
     1019 
     1020Can inspect websites' pages for whether it's relevant/auto-translated as follows: 
     1021    db.getCollection('Webpages').find({URL:/svenkirsten.com/, mriSentenceCount: {$gt: 0}}) 
     1022 
    9281023 
    9291024* CN: Only 1/113 sites from CN stood out as being of interest: http://kiwi2china.com/ 
     
    9371032    http://splaf.free.fr/pfurb.html - Tahiti, French Polynesian, ... island names 
    9381033    http://mi.fitnessrebates.com - Uses https://wordpress.org/plugins/weglot/ wordpress-compatible multilingual plugin, which ensures translated pages get indexed by google - exactly what we want to avoid 
    939 *  
    940  
     1034* 
     1035 
     1036 
     1037DE: 
     1038http://etymologie.info/~e/n_/nz-___reg.html - placenames, not meaningful 
     1039!! https://www.cartogiraffe.com/ - some genuine pages (Rarotongan), but one page is in Czech that had a single word misindentified as MRI 
     1040~ http://svenkirsten.com/ - one page mentions "tiki" but the rest is in English. The other is an (English) caption of "Book of Tiki A Maori Maiden" 
     1041- herocity - autotranslated 
     1042- weltderberge.de - 3 pages mention NZ mountains by name. 
     1043~ (arts.mythologica.fr) https://mythologica.fr/oceanie/texte/pantheon_polynesien.pdf - mentions certain Maori Gods and other Polynesian Gods by name. 
     1044- https://traynews.com - nothing in MRI, misdetected 
     1045~ http://klaaskoehne.de/galleries/nzl-taranaki/index.html - mentions NZ mountain names 
     1046- http://www.nierstrasz.org/deGrauwRegister.rtf - misdetected European (Dutch) names as MRI 
     1047- https://afrikhepri.org/mi/ - autotranslated 
     1048- https://www.tvteile.de - pure German pages, misdetected "Automatik" as a Maori language word 
     1049- etoile-de-lune.net - 5 pages containing 1 sentence each but none with 2 sentences detected 
     1050 
     1051- ITALY: 
     1052  http://oipaz.net/IMG/GalleriaAotearoa/ - NZ photogallery with each photo captioned by placename 
     1053  http://www.marcosanti.it/Reportage/Oceania_ph/Nuova_Zelanda/ - each photo captioned by NZ placename 
     1054  http://www.pegasoesmicamion.com/ - REO abbreviation misidentified, also in REO%20PUBLICIDAD.htm 
     1055- AUSTRIA: 
     1056  petit-prince.at - Tahitian and Wayuu (Venezuela) translations of Le Petit Prince 
     1057  http://www.tmtmm.net/newzealand - photos from NZ named after places and people's names 
     1058- ROMANIA: parohiauceadesus.ro - Sentences of single Romanian words misidentified. 
     1059- ISRAEL: 
     1060  http://www.daat.ac.il - misidentification of "no." as MRI, and Hebrew words. 
     1061  https://www.hitiaotera.com/ - misidentifiation of Tahitian pages 
     1062- RUSSIA: https://www.gismeteo.lv - misidentification of an email address 
     1063- JAPAN: http://yutaka.it-n.jp - many pages of scientific names of (plants?) which are often misdetected as MRI 
     1064!! Ireland, ie: https://coggle.it 
     1065- IRAN: https://www.dideo.ir/v/yt/d6cgya0ze-E - video title from MaoriTelevision website 
     1066? - CZECH republic: https://www.fipojobs.com/new-zealand/jobs-work-p-1 - NZ job position title in MRI but rest in English 
     1067- SPAIN: http://www.info-hoteles.com/nz/2/hotels_lake_rotoiti.asp - 2 uses of the same placename 
     1068- SINGAPORE: https://omg-solutions.com - autotranslated 
     1069- TURKEY: https://www.elitedeluxe.com.tr/mi/yatak-odasi-takimlari - autotranslated 
     1070- MEXICO: http://www.gelbukh.com - misidentification, lines of just numbers or phrases like "Area Chair" in English and Russian CVs. 
     1071- FINLAND: http://pertti.com - travelogue, placenames 
     1072- SWITZERLAND CH: 
     1073  nicoledidi.ch - blog, placenames 
     1074  https://photos.axelebert.org - Tahiti related content 
     1075- UNKNOWN: https://www.viveipcl.com: tours website, placenames mentioned 
     1076#- EU: https://www.the-good-stuff-factory.be/mi/ : Autotranslated 
     1077!! - BULGARIA: http://anitra.net/activism/humanrights/UDHR/rrt_print.htm (2 pages) 
     1078 
     1079 
     1080TREATING AUSTRALIA AND GREAT BRITAIN MORE SPECIALLY (don't ignore /mi in URL, same as with NZ, but do leave out .nz TLDs since we cover them under NZ - TODO: later find country codes of all .nz TLDs): 
     1081[nothing found under "UK", only under "GB"] 
     1082 
     1083db.getCollection('Websites').find({ 
     1084    domain: {$not: /.nz$/}, 
     1085    numPagesContainingMRI: {$gt: 0},     
     1086    $or: [{geoLocationCountryCode: "AU"}, {geoLocationCountryCode: "GB"}] 
     1087}).count() 
     108811 
     1089 
     1090db.Websites.aggregate([ 
     1091    { 
     1092        $match: { 
     1093           domain: {$not: /.nz$/}, 
     1094           numPagesContainingMRI: {$gt: 0},            
     1095           $or: [{geoLocationCountryCode: "AU"}, {geoLocationCountryCode: "GB"}] 
     1096        } 
     1097    }, 
     1098    { $unwind: "$geoLocationCountryCode" }, 
     1099    { 
     1100        $group: { 
     1101            _id: {$toLower: '$geoLocationCountryCode'}, 
     1102            count: { $sum: 1 }, 
     1103            domain: { $addToSet: '$domain' } 
     1104        } 
     1105    }, 
     1106    { $sort : { count : -1} } 
     1107]); 
     1108 
     1109AUSTRALIA: 
     1110!!  https://www.kiwiproperty.com - e.g. https://www.kiwiproperty.com/the-base/mi/he-paepaki/ has some actual MRI sentences. [Not autotranslated] 
     1111?   http://fionajack.net - Wellington gallery of artist. A few occurrences of Kia Ora in a title like context (e.g. "Street Party Kia Ora! Kia Ora!") 
     1112!!  https://infogram.com/te-marautanga-o-aotearoa-moe-pld-allocations-2012-1go502ygvn562jd  - site of individual pages (like docs.google.com). This one has a relevant infogram image. 
     1113!!  https://koreromaori.com - some actual Maori language sentences 
     1114    http://theunderwaterworld.com/Galleries/Roimata/roimata-frame.html - placenames 
     1115 
     1116UK: 
     1117    http://www.wordsearchfun.com/200628_Word_Find_wordsearch.html - 2 word games with Maori words (one of them has 3 different views, e.g. print view) 
     1118?   https://omniatlas.com/maps/australasia/18400206/plain/ - historical map with Maori iwi names over NZ map regions 
     1119?      https://omniatlas.com/maps/australasia/18400206/ - historical map of Australia and NZ at the time of the Treaty of Waitangi, with events marked in English 
     1120    https://centrallanguageschool.com - AUTOTRANSLATED 
     1121    https://www.solasolv.com - Autotranslated product site 
     1122    http://mikestephens.co.uk/ - photo captions containing NZ placenames 
     1123    http://www.woolrych.org/nzholiday2004/ - photogallery captioned with NZ placenames 
    9411124-------------- 
    9421125 
     
    9591142 
    9601143--------------------- 
     1144 
     1145Count of websites that have at least 1 page containing at least one sentence detected as MRI 
     1146AND which websites have mi in the URL path: 
     1147 
     1148db.getCollection('Websites').find({$and: [{numPagesContainingMRI: {$gt: 0}},{urlContainsLangCodeInPath: true}]}).count() 
     1149 
     1150491 
     1151 
     1152 
     1153 
     1154# The websites that have some MRI detected AND which are either in NZ or with NZ TLD 
     1155# or (so if they're from overseas) don't contain /mi or mi.* in URL path: 
     1156 
     1157db.getCollection('Websites').find({$and: [{numPagesContainingMRI: {$gt: 0}},{$or: [{geoLocationCountryCode: "NZ"}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]}]}).count() 
     1158396 
     1159 
     1160Include Australia (to get the valid "kiwiproperty.com" website included in the result list): 
     1161 
     1162db.getCollection('Websites').find({$and: [ 
     1163                {numPagesContainingMRI: {$gt: 0}}, 
     1164                {$or: [{geoLocationCountryCode: /(NZ|AU)/}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]} 
     1165            ]}).count() 
     1166 
     1167397 
     1168 
     1169# aggregate results by a count of country codes 
     1170db.Websites.aggregate([ 
     1171    { 
     1172        $match: { 
     1173            $and: [ 
     1174                {numPagesContainingMRI: {$gt: 0}}, 
     1175                {$or: [{geoLocationCountryCode: /(NZ|AU)/}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]} 
     1176            ] 
     1177        } 
     1178    }, 
     1179    { $unwind: "$geoLocationCountryCode" }, 
     1180    { 
     1181        $group: { 
     1182            _id: {$toLower: '$geoLocationCountryCode'}, 
     1183            count: { $sum: 1 } 
     1184        } 
     1185    }, 
     1186    { $sort : { count : -1} } 
     1187]); 
     1188 
     1189 
     1190# Just considering those sites outside NZ or not with .nz TLD: 
     1191 
     1192db.getCollection('Websites').find({$and: [ 
     1193                {geoLocationCountryCode: {$ne: "NZ"}}, 
     1194                {domain: {$not: /\.nz/}}, 
     1195                {numPagesContainingMRI: {$gt: 0}}, 
     1196                {$or: [{geoLocationCountryCode: "AU"}, {urlContainsLangCodeInPath: false}]}             
     1197            ]}).count() 
     1198 
     1199221 websites 
     1200 
     1201# counts by country code excluding NZ related sites 
     1202db.Websites.aggregate([ 
     1203    { 
     1204        $match: { 
     1205            $and: [ 
     1206                {geoLocationCountryCode: {$ne: "NZ"}}, 
     1207                {domain: {$not: /\.nz/}}, 
     1208                {numPagesContainingMRI: {$gt: 0}}, 
     1209                {$or: [{geoLocationCountryCode: "AU"}, {urlContainsLangCodeInPath: false}]}             
     1210            ] 
     1211        } 
     1212    }, 
     1213    { $unwind: "$geoLocationCountryCode" }, 
     1214    { 
     1215        $group: { 
     1216            _id: {$toLower: '$geoLocationCountryCode'}, 
     1217            count: { $sum: 1 }, 
     1218            domain: { $addToSet: '$domain' } 
     1219        } 
     1220    }, 
     1221    { $sort : { count : -1} } 
     1222]); 
     1223 
     1224 
     1225# But to produce the tentative non-product sites, we also want the aggregate for all NZ sites (from NZ or with .nz tld): 
     1226db.getCollection('Websites').find({$and: [ 
     1227                {numPagesContainingMRI: {$gt: 0}}, 
     1228                {$or: [{geoLocationCountryCode:"NZ"},{domain: /\.nz/}]} 
     1229            ]}).count() 
     1230 
     1231176 
     1232 
     1233(Total is 221+176 = 397, which adds up). 
     1234 
     1235# Get the count (and domain listing) output put under a hardcoded  _id of "nz": 
     1236db.Websites.aggregate([ 
     1237    { 
     1238        $match: { 
     1239            $and: [ 
     1240                {numPagesContainingMRI: {$gt: 0}}, 
     1241                {$or: [{geoLocationCountryCode:"NZ"},{domain: /\.nz/}]} 
     1242            ] 
     1243        } 
     1244    }, 
     1245    { $unwind: "$geoLocationCountryCode" }, 
     1246    { 
     1247        $group: { 
     1248            _id: "nz", 
     1249            count: { $sum: 1 }, 
     1250            domain: { $addToSet: '$domain' } 
     1251        } 
     1252    }, 
     1253    { $sort : { count : -1} } 
     1254]); 
  • other-projects/maori-lang-detection/mongodb-data/counts_allCrawledSites.json

    r33802 r33813  
     1/* 
     2Num websites: 
     3db.getCollection('Websites').find({}).count() 
     4= 1445  
     5 
     6Num webpages 
     7db.getCollection('Webpages').find({}).count() 
     8= 117496 
     9 
     10Count of country codes for all sites: 
     11db.Websites.aggregate([ 
     12    
     13    { $unwind: "$geoLocationCountryCode" }, 
     14    { 
     15        $group: { 
     16            _id: "$geoLocationCountryCode", 
     17            count: { $sum: 1 } 
     18        } 
     19    }, 
     20    { $sort : { count : -1} } 
     21]); 
     22*/ 
     23 
    124/* 1 */ 
    225{ 
    326    "_id" : "US", 
    4     "count" : 686.0 
     27    "count" : 696.0 
    528} 
    629 
     
    831{ 
    932    "_id" : "UNKNOWN", 
    10     "count" : 171.0 
     33    "count" : 173.0 
    1134} 
    1235 
    1336/* 3 */ 
    1437{ 
     38    "_id" : "CN", 
     39    "count" : 125.0 
     40} 
     41 
     42/* 4 */ 
     43{ 
    1544    "_id" : "NZ", 
    16     "count" : 128.0 
    17 } 
    18  
    19 /* 4 */ 
    20 { 
    21     "_id" : "CN", 
    22     "count" : 124.0 
     45    "count" : 115.0 
    2346} 
    2447 
     
    4467{ 
    4568    "_id" : "NL", 
    46     "count" : 31.0 
     69    "count" : 32.0 
    4770} 
    4871 
     
    5073{ 
    5174    "_id" : "CA", 
    52     "count" : 21.0 
     75    "count" : 19.0 
    5376} 
    5477 
     
    6184/* 11 */ 
    6285{ 
     86    "_id" : "DK", 
     87    "count" : 10.0 
     88} 
     89 
     90/* 12 */ 
     91{ 
    6392    "_id" : "JP", 
    6493    "count" : 10.0 
    6594} 
    6695 
    67 /* 12 */ 
    68 { 
    69     "_id" : "DK", 
    70     "count" : 10.0 
    71 } 
    72  
    7396/* 13 */ 
    7497{ 
     
    79102/* 14 */ 
    80103{ 
     104    "_id" : "RU", 
     105    "count" : 7.0 
     106} 
     107 
     108/* 15 */ 
     109{ 
     110    "_id" : "HK", 
     111    "count" : 7.0 
     112} 
     113 
     114/* 16 */ 
     115{ 
    81116    "_id" : "CZ", 
    82117    "count" : 7.0 
    83118} 
    84119 
    85 /* 15 */ 
    86 { 
    87     "_id" : "RU", 
    88     "count" : 7.0 
    89 } 
    90  
    91 /* 16 */ 
    92 { 
    93     "_id" : "HK", 
    94     "count" : 7.0 
    95 } 
    96  
    97120/* 17 */ 
    98 { 
    99     "_id" : "IT", 
    100     "count" : 5.0 
    101 } 
    102  
    103 /* 18 */ 
    104 { 
    105     "_id" : "SE", 
    106     "count" : 5.0 
    107 } 
    108  
    109 /* 19 */ 
    110121{ 
    111122    "_id" : "UA", 
     
    113124} 
    114125 
    115 /* 20 */ 
     126/* 18 */ 
    116127{ 
    117128    "_id" : "IE", 
     
    119130} 
    120131 
     132/* 19 */ 
     133{ 
     134    "_id" : "SE", 
     135    "count" : 4.0 
     136} 
     137 
     138/* 20 */ 
     139{ 
     140    "_id" : "IT", 
     141    "count" : 4.0 
     142} 
     143 
    121144/* 21 */ 
    122145{ 
     146    "_id" : "RO", 
     147    "count" : 3.0 
     148} 
     149 
     150/* 22 */ 
     151{ 
     152    "_id" : "SG", 
     153    "count" : 3.0 
     154} 
     155 
     156/* 23 */ 
     157{ 
     158    "_id" : "AT", 
     159    "count" : 3.0 
     160} 
     161 
     162/* 24 */ 
     163{ 
     164    "_id" : "CH", 
     165    "count" : 3.0 
     166} 
     167 
     168/* 25 */ 
     169{ 
     170    "_id" : "IL", 
     171    "count" : 3.0 
     172} 
     173 
     174/* 26 */ 
     175{ 
    123176    "_id" : "IN", 
    124177    "count" : 3.0 
    125178} 
    126179 
    127 /* 22 */ 
    128 { 
    129     "_id" : "CH", 
    130     "count" : 3.0 
    131 } 
    132  
    133 /* 23 */ 
    134 { 
    135     "_id" : "IL", 
    136     "count" : 3.0 
    137 } 
    138  
    139 /* 24 */ 
    140 { 
    141     "_id" : "AT", 
    142     "count" : 3.0 
    143 } 
    144  
    145 /* 25 */ 
    146 { 
    147     "_id" : "RO", 
    148     "count" : 2.0 
    149 } 
    150  
    151 /* 26 */ 
     180/* 27 */ 
     181{ 
     182    "_id" : "PL", 
     183    "count" : 2.0 
     184} 
     185 
     186/* 28 */ 
     187{ 
     188    "_id" : "ZA", 
     189    "count" : 2.0 
     190} 
     191 
     192/* 29 */ 
     193{ 
     194    "_id" : "VG", 
     195    "count" : 2.0 
     196} 
     197 
     198/* 30 */ 
    152199{ 
    153200    "_id" : "CK", 
     
    155202} 
    156203 
    157 /* 27 */ 
    158 { 
    159     "_id" : "VG", 
    160     "count" : 2.0 
    161 } 
    162  
    163 /* 28 */ 
    164 { 
    165     "_id" : "ZA", 
    166     "count" : 2.0 
    167 } 
    168  
    169 /* 29 */ 
    170 { 
    171     "_id" : "PL", 
    172     "count" : 2.0 
    173 } 
    174  
    175 /* 30 */ 
     204/* 31 */ 
     205{ 
     206    "_id" : "BG", 
     207    "count" : 1.0 
     208} 
     209 
     210/* 32 */ 
     211{ 
     212    "_id" : "PF", 
     213    "count" : 1.0 
     214} 
     215 
     216/* 33 */ 
     217{ 
     218    "_id" : "IO", 
     219    "count" : 1.0 
     220} 
     221 
     222/* 34 */ 
    176223{ 
    177224    "_id" : "GR", 
     
    179226} 
    180227 
    181 /* 31 */ 
     228/* 35 */ 
     229{ 
     230    "_id" : "MX", 
     231    "count" : 1.0 
     232} 
     233 
     234/* 36 */ 
     235{ 
     236    "_id" : "TR", 
     237    "count" : 1.0 
     238} 
     239 
     240/* 37 */ 
     241{ 
     242    "_id" : "ME", 
     243    "count" : 1.0 
     244} 
     245 
     246/* 38 */ 
    182247{ 
    183248    "_id" : "FI", 
     
    185250} 
    186251 
    187 /* 32 */ 
    188 { 
    189     "_id" : "ME", 
    190     "count" : 1.0 
    191 } 
    192  
    193 /* 33 */ 
     252/* 39 */ 
     253{ 
     254    "_id" : "EU", 
     255    "count" : 1.0 
     256} 
     257 
     258/* 40 */ 
     259{ 
     260    "_id" : "IR", 
     261    "count" : 1.0 
     262} 
     263 
     264/* 41 */ 
    194265{ 
    195266    "_id" : "PT", 
    196267    "count" : 1.0 
    197268} 
    198  
    199 /* 34 */ 
    200 { 
    201     "_id" : "IR", 
    202     "count" : 1.0 
    203 } 
    204  
    205 /* 35 */ 
    206 { 
    207     "_id" : "EU", 
    208     "count" : 1.0 
    209 } 
    210  
    211 /* 36 */ 
    212 { 
    213     "_id" : "TR", 
    214     "count" : 1.0 
    215 } 
    216  
    217 /* 37 */ 
    218 { 
    219     "_id" : "MX", 
    220     "count" : 1.0 
    221 } 
    222  
    223 /* 38 */ 
    224 { 
    225     "_id" : "IO", 
    226     "count" : 1.0 
    227 } 
    228  
    229 /* 39 */ 
    230 { 
    231     "_id" : "PF", 
    232     "count" : 1.0 
    233 } 
    234  
    235 /* 40 */ 
    236 { 
    237     "_id" : "BG", 
    238     "count" : 1.0 
    239 } 
    240  
    241 /* 41 */ 
    242 { 
    243     "_id" : "SG", 
    244     "count" : 1.0 
    245 } 
  • other-projects/maori-lang-detection/mongodb-data/counts_miInUrlPath.json

    r33797 r33813  
    22{ 
    33    "_id" : "us", 
    4     "count" : 93.0 
     4    "count" : 408.0 
    55} 
    66 
     
    88{ 
    99    "_id" : "cn", 
     10    "count" : 123.0 
     11} 
     12 
     13/* 3 */ 
     14{ 
     15    "_id" : "fr", 
     16    "count" : 34.0 
     17} 
     18 
     19/* 4 */ 
     20{ 
     21    "_id" : "unknown", 
     22    "count" : 19.0 
     23} 
     24 
     25/* 5 */ 
     26{ 
     27    "_id" : "nz", 
     28    "count" : 14.0 
     29} 
     30 
     31/* 6 */ 
     32{ 
     33    "_id" : "de", 
     34    "count" : 12.0 
     35} 
     36 
     37/* 7 */ 
     38{ 
     39    "_id" : "nl", 
     40    "count" : 8.0 
     41} 
     42 
     43/* 8 */ 
     44{ 
     45    "_id" : "hk", 
    1046    "count" : 7.0 
    1147} 
    1248 
    13 /* 3 */ 
     49/* 9 */ 
    1450{ 
    1551    "_id" : "au", 
     
    1753} 
    1854 
    19 /* 4 */ 
     55/* 10 */ 
    2056{ 
    21     "_id" : "de", 
     57    "_id" : "ca", 
     58    "count" : 7.0 
     59} 
     60 
     61/* 11 */ 
     62{ 
     63    "_id" : "gb", 
    2264    "count" : 5.0 
    2365} 
    2466 
    25 /* 5 */ 
     67/* 12 */ 
    2668{ 
    27     "_id" : "hk", 
     69    "_id" : "jp", 
    2870    "count" : 5.0 
    2971} 
    3072 
    31 /* 6 */ 
     73/* 13 */ 
    3274{ 
    33     "_id" : "nz", 
    34     "count" : 5.0 
     75    "_id" : "ua", 
     76    "count" : 4.0 
    3577} 
    3678 
    37 /* 7 */ 
     79/* 14 */ 
    3880{ 
    3981    "_id" : "ru", 
     
    4183} 
    4284 
    43 /* 8 */ 
     85/* 15 */ 
    4486{ 
    45     "_id" : "gb", 
    46     "count" : 3.0 
     87    "_id" : "vg", 
     88    "count" : 2.0 
    4789} 
    4890 
    49 /* 9 */ 
     91/* 16 */ 
    5092{ 
    51     "_id" : "jp", 
    52     "count" : 3.0 
     93    "_id" : "sg", 
     94    "count" : 2.0 
    5395} 
    5496 
    55 /* 10 */ 
     97/* 17 */ 
    5698{ 
    5799    "_id" : "dk", 
     
    59101} 
    60102 
    61 /* 11 */ 
     103/* 18 */ 
    62104{ 
    63     "_id" : "ca", 
    64     "count" : 2.0 
    65 } 
    66  
    67 /* 12 */ 
    68 { 
    69     "_id" : "fr", 
    70     "count" : 2.0 
    71 } 
    72  
    73 /* 13 */ 
    74 { 
    75     "_id" : "vg", 
    76     "count" : 2.0 
    77 } 
    78  
    79 /* 14 */ 
    80 { 
    81     "_id" : "nl", 
     105    "_id" : "se", 
    82106    "count" : 1.0 
    83107} 
    84108 
    85 /* 15 */ 
     109/* 19 */ 
    86110{ 
    87     "_id" : "ch", 
     111    "_id" : "ie", 
    88112    "count" : 1.0 
    89113} 
    90114 
    91 /* 16 */ 
     115/* 20 */ 
     116{ 
     117    "_id" : "es", 
     118    "count" : 1.0 
     119} 
     120 
     121/* 21 */ 
     122{ 
     123    "_id" : "tr", 
     124    "count" : 1.0 
     125} 
     126 
     127/* 22 */ 
    92128{ 
    93129    "_id" : "za", 
     
    95131} 
    96132 
    97 /* 17 */ 
     133/* 23 */ 
    98134{ 
    99135    "_id" : "cz", 
     
    101137} 
    102138 
    103 /* 18 */ 
     139/* 24 */ 
    104140{ 
    105     "_id" : "kr", 
     141    "_id" : "eu", 
    106142    "count" : 1.0 
    107143} 
    108  
    109 /* 19 */ 
    110 { 
    111     "_id" : "ua", 
    112     "count" : 1.0 
    113 } 
  • other-projects/maori-lang-detection/mongodb-data/counts_noMiInUrlPath.json

    r33798 r33813  
    22{ 
    33    "_id" : "us", 
    4     "count" : 612.0 
     4    "count" : 288.0 
    55} 
    66 
     
    88{ 
    99    "_id" : "unknown", 
    10     "count" : 150.0 
     10    "count" : 154.0 
    1111} 
    1212 
     
    1414{ 
    1515    "_id" : "nz", 
    16     "count" : 112.0 
     16    "count" : 101.0 
    1717} 
    1818 
    1919/* 4 */ 
    2020{ 
    21     "_id" : "cn", 
    22     "count" : 106.0 
     21    "_id" : "de", 
     22    "count" : 40.0 
    2323} 
    2424 
    2525/* 5 */ 
    2626{ 
     27    "_id" : "au", 
     28    "count" : 36.0 
     29} 
     30 
     31/* 6 */ 
     32{ 
    2733    "_id" : "fr", 
    28     "count" : 69.0 
    29 } 
    30  
    31 /* 6 */ 
    32 { 
    33     "_id" : "de", 
    34     "count" : 46.0 
     34    "count" : 35.0 
    3535} 
    3636 
    3737/* 7 */ 
    3838{ 
    39     "_id" : "au", 
    40     "count" : 35.0 
     39    "_id" : "nl", 
     40    "count" : 24.0 
    4141} 
    4242 
    4343/* 8 */ 
    4444{ 
    45     "_id" : "nl", 
    46     "count" : 30.0 
     45    "_id" : "gb", 
     46    "count" : 13.0 
    4747} 
    4848 
     
    5050{ 
    5151    "_id" : "ca", 
    52     "count" : 21.0 
     52    "count" : 12.0 
    5353} 
    5454 
    5555/* 10 */ 
    56 { 
    57     "_id" : "gb", 
    58     "count" : 15.0 
    59 } 
    60  
    61 /* 11 */ 
    62 { 
    63     "_id" : "es", 
    64     "count" : 8.0 
    65 } 
    66  
    67 /* 12 */ 
    6856{ 
    6957    "_id" : "dk", 
     
    7159} 
    7260 
     61/* 11 */ 
     62{ 
     63    "_id" : "es", 
     64    "count" : 7.0 
     65} 
     66 
     67/* 12 */ 
     68{ 
     69    "_id" : "cz", 
     70    "count" : 6.0 
     71} 
     72 
    7373/* 13 */ 
    7474{ 
    75     "_id" : "cz", 
    76     "count" : 7.0 
     75    "_id" : "jp", 
     76    "count" : 5.0 
    7777} 
    7878 
    7979/* 14 */ 
    80 { 
    81     "_id" : "jp", 
    82     "count" : 7.0 
    83 } 
    84  
    85 /* 15 */ 
    8680{ 
    8781    "_id" : "it", 
     
    8983} 
    9084 
     85/* 15 */ 
     86{ 
     87    "_id" : "in", 
     88    "count" : 3.0 
     89} 
     90 
    9191/* 16 */ 
    9292{ 
    93     "_id" : "in", 
     93    "_id" : "ie", 
    9494    "count" : 3.0 
    9595} 
     
    9797/* 17 */ 
    9898{ 
    99     "_id" : "ua", 
     99    "_id" : "at", 
    100100    "count" : 3.0 
    101101} 
     
    103103/* 18 */ 
    104104{ 
    105     "_id" : "at", 
     105    "_id" : "ro", 
    106106    "count" : 3.0 
    107107} 
     
    109109/* 19 */ 
    110110{ 
    111     "_id" : "ie", 
     111    "_id" : "ru", 
    112112    "count" : 3.0 
    113113} 
     
    115115/* 20 */ 
    116116{ 
     117    "_id" : "se", 
     118    "count" : 3.0 
     119} 
     120 
     121/* 21 */ 
     122{ 
    117123    "_id" : "ch", 
    118124    "count" : 3.0 
    119125} 
    120126 
    121 /* 21 */ 
    122 { 
    123     "_id" : "se", 
    124     "count" : 3.0 
    125 } 
    126  
    127127/* 22 */ 
    128128{ 
    129     "_id" : "ru", 
     129    "_id" : "il", 
    130130    "count" : 3.0 
    131131} 
     
    133133/* 23 */ 
    134134{ 
    135     "_id" : "il", 
    136     "count" : 3.0 
     135    "_id" : "pl", 
     136    "count" : 2.0 
    137137} 
    138138 
     
    145145/* 25 */ 
    146146{ 
    147     "_id" : "pl", 
     147    "_id" : "cn", 
    148148    "count" : 2.0 
    149149} 
     
    151151/* 26 */ 
    152152{ 
    153     "_id" : "ro", 
    154     "count" : 2.0 
     153    "_id" : "sg", 
     154    "count" : 1.0 
    155155} 
    156156 
    157157/* 27 */ 
    158158{ 
    159     "_id" : "sg", 
     159    "_id" : "mx", 
    160160    "count" : 1.0 
    161161} 
     
    163163/* 28 */ 
    164164{ 
    165     "_id" : "mx", 
     165    "_id" : "io", 
    166166    "count" : 1.0 
    167167} 
     
    169169/* 29 */ 
    170170{ 
     171    "_id" : "za", 
     172    "count" : 1.0 
     173} 
     174 
     175/* 30 */ 
     176{ 
    171177    "_id" : "ir", 
    172178    "count" : 1.0 
    173179} 
    174180 
    175 /* 30 */ 
    176 { 
    177     "_id" : "eu", 
    178     "count" : 1.0 
    179 } 
    180  
    181181/* 31 */ 
    182182{ 
     183    "_id" : "ua", 
     184    "count" : 1.0 
     185} 
     186 
     187/* 32 */ 
     188{ 
    183189    "_id" : "bg", 
    184190    "count" : 1.0 
    185191} 
    186192 
    187 /* 32 */ 
     193/* 33 */ 
    188194{ 
    189195    "_id" : "gr", 
     
    191197} 
    192198 
    193 /* 33 */ 
     199/* 34 */ 
     200{ 
     201    "_id" : "fi", 
     202    "count" : 1.0 
     203} 
     204 
     205/* 35 */ 
     206{ 
     207    "_id" : "me", 
     208    "count" : 1.0 
     209} 
     210 
     211/* 36 */ 
     212{ 
     213    "_id" : "pf", 
     214    "count" : 1.0 
     215} 
     216 
     217/* 37 */ 
    194218{ 
    195219    "_id" : "pt", 
     
    197221} 
    198222 
    199 /* 34 */ 
    200 { 
    201     "_id" : "fi", 
    202     "count" : 1.0 
    203 } 
    204  
    205 /* 35 */ 
    206 { 
    207     "_id" : "pf", 
    208     "count" : 1.0 
    209 } 
    210  
  • other-projects/maori-lang-detection/mongodb-data/counts_sitesWithPagesContainingMRI.json

    r33803 r33813  
     1/* 
     2Number of sites containing at least one sentence for which OpenNLP detected the best language = MRI 
     3 
     4db.getCollection('Websites').find({numPagesContainingMRI: {$gt: 0}}).count() 
     5868 
     6 
     7 
     8Obviously, the following should be equal to that: 
     9 
     10db.getCollection('Websites').find({ $or: [ { numPagesInMRI: { $gt: 0 } }, { numPagesContainingMRI: {$gt: 0} } ] } ).count() 
     11868 
     12 
     13 
     14Count of country codes for sites that have at least one page containing at least one sentence detected as MRI by OpenNLP: 
     15 
     16db.Websites.aggregate([ 
     17    { 
     18        $match: { 
     19            numPagesContainingMRI: {$gt: 0} 
     20        } 
     21    }, 
     22    { $unwind: "$geoLocationCountryCode" }, 
     23    { 
     24        $group: { 
     25            _id: {$toLower: '$geoLocationCountryCode'}, 
     26            count: { $sum: 1 } 
     27        } 
     28    }, 
     29    { $sort : { count : -1} } 
     30]); 
     31 
     32*/ 
     33 
    134/* 1 */ 
    235{ 
    336    "_id" : "us", 
    4     "count" : 479.0 
     37    "count" : 486.0 
    538} 
    639 
     
    1447{ 
    1548    "_id" : "nz", 
    16     "count" : 98.0 
     49    "count" : 89.0 
    1750} 
    1851 
     
    3871{ 
    3972    "_id" : "au", 
    40     "count" : 20.0 
     73    "count" : 21.0 
    4174} 
    4275 
     
    4477{ 
    4578    "_id" : "ca", 
    46     "count" : 13.0 
     79    "count" : 12.0 
    4780} 
    4881 
     
    5588/* 10 */ 
    5689{ 
     90    "_id" : "es", 
     91    "count" : 7.0 
     92} 
     93 
     94/* 11 */ 
     95{ 
    5796    "_id" : "gb", 
    5897    "count" : 7.0 
    5998} 
    6099 
    61 /* 11 */ 
    62 { 
    63     "_id" : "es", 
    64     "count" : 7.0 
    65 } 
    66  
    67100/* 12 */ 
    68101{ 
     
    73106/* 13 */ 
    74107{ 
     108    "_id" : "unknown", 
     109    "count" : 3.0 
     110} 
     111 
     112/* 14 */ 
     113{ 
     114    "_id" : "at", 
     115    "count" : 3.0 
     116} 
     117 
     118/* 15 */ 
     119{ 
     120    "_id" : "ro", 
     121    "count" : 3.0 
     122} 
     123 
     124/* 16 */ 
     125{ 
    75126    "_id" : "it", 
    76127    "count" : 3.0 
    77128} 
    78129 
    79 /* 14 */ 
    80 { 
    81     "_id" : "at", 
    82     "count" : 3.0 
    83 } 
    84  
    85 /* 15 */ 
     130/* 17 */ 
     131{ 
     132    "_id" : "sg", 
     133    "count" : 2.0 
     134} 
     135 
     136/* 18 */ 
     137{ 
     138    "_id" : "jp", 
     139    "count" : 2.0 
     140} 
     141 
     142/* 19 */ 
     143{ 
     144    "_id" : "ie", 
     145    "count" : 2.0 
     146} 
     147 
     148/* 20 */ 
     149{ 
     150    "_id" : "hk", 
     151    "count" : 2.0 
     152} 
     153 
     154/* 21 */ 
     155{ 
     156    "_id" : "ua", 
     157    "count" : 2.0 
     158} 
     159 
     160/* 22 */ 
     161{ 
     162    "_id" : "ru", 
     163    "count" : 2.0 
     164} 
     165 
     166/* 23 */ 
     167{ 
     168    "_id" : "ch", 
     169    "count" : 2.0 
     170} 
     171 
     172/* 24 */ 
    86173{ 
    87174    "_id" : "il", 
     
    89176} 
    90177 
    91 /* 16 */ 
    92 { 
    93     "_id" : "ch", 
    94     "count" : 2.0 
    95 } 
    96  
    97 /* 17 */ 
    98 { 
    99     "_id" : "ro", 
    100     "count" : 2.0 
    101 } 
    102  
    103 /* 18 */ 
    104 { 
    105     "_id" : "ru", 
    106     "count" : 2.0 
    107 } 
    108  
    109 /* 19 */ 
    110 { 
    111     "_id" : "ie", 
    112     "count" : 2.0 
    113 } 
    114  
    115 /* 20 */ 
    116 { 
    117     "_id" : "jp", 
    118     "count" : 2.0 
    119 } 
    120  
    121 /* 21 */ 
    122 { 
    123     "_id" : "hk", 
    124     "count" : 2.0 
    125 } 
    126  
    127 /* 22 */ 
    128 { 
    129     "_id" : "ua", 
    130     "count" : 2.0 
    131 } 
    132  
    133 /* 23 */ 
    134 { 
    135     "_id" : "unknown", 
    136     "count" : 2.0 
    137 } 
    138  
    139 /* 24 */ 
     178/* 25 */ 
     179{ 
     180    "_id" : "tr", 
     181    "count" : 1.0 
     182} 
     183 
     184/* 26 */ 
    140185{ 
    141186    "_id" : "mx", 
     
    143188} 
    144189 
    145 /* 25 */ 
     190/* 27 */ 
     191{ 
     192    "_id" : "ir", 
     193    "count" : 1.0 
     194} 
     195 
     196/* 28 */ 
     197{ 
     198    "_id" : "gr", 
     199    "count" : 1.0 
     200} 
     201 
     202/* 29 */ 
     203{ 
     204    "_id" : "bg", 
     205    "count" : 1.0 
     206} 
     207 
     208/* 30 */ 
     209{ 
     210    "_id" : "eu", 
     211    "count" : 1.0 
     212} 
     213 
     214/* 31 */ 
    146215{ 
    147216    "_id" : "fi", 
    148217    "count" : 1.0 
    149218} 
    150  
    151 /* 26 */ 
    152 { 
    153     "_id" : "se", 
    154     "count" : 1.0 
    155 } 
    156  
    157 /* 27 */ 
    158 { 
    159     "_id" : "gr", 
    160     "count" : 1.0 
    161 } 
    162  
    163 /* 28 */ 
    164 { 
    165     "_id" : "eu", 
    166     "count" : 1.0 
    167 } 
    168  
    169 /* 29 */ 
    170 { 
    171     "_id" : "bg", 
    172     "count" : 1.0 
    173 } 
    174  
    175 /* 30 */ 
    176 { 
    177     "_id" : "ir", 
    178     "count" : 1.0 
    179 } 
    180  
    181 /* 31 */ 
    182 { 
    183     "_id" : "tr", 
    184     "count" : 1.0 
    185 } 
    186  
    187 /* 32 */ 
    188 { 
    189     "_id" : "sg", 
    190     "count" : 1.0 
    191 } 
  • other-projects/maori-lang-detection/mongodb-data/counts_sitesWithPagesInMRI.json

    r33805 r33813  
     1/* 
     2Number of websites that have 1 or more pages detected as being in Maori, a positive numPagesInMRI. 
     3 
     4db.getCollection('Websites').find({numPagesInMRI: { $gt: 0}}).count() 
     5= 361 
     6 
     7Count of country codes for sites that have at least one page detected as MRI: 
     8 
     9db.Websites.aggregate([ 
     10    { 
     11        $match: { 
     12            numPagesInMRI: {$gt: 0} 
     13        } 
     14    }, 
     15    { $unwind: "$geoLocationCountryCode" }, 
     16    { 
     17        $group: { 
     18            _id: {$toLower: '$geoLocationCountryCode'}, 
     19            count: { $sum: 1 } 
     20        } 
     21    }, 
     22    { $sort : { count : -1} } 
     23]); 
     24*/ 
     25 
    126/* 1 */ 
    227{ 
    328    "_id" : "us", 
    4     "count" : 207.0 
     29    "count" : 206.0 
    530} 
    631 
     
    4974/* 9 */ 
    5075{ 
    51     "_id" : "ca", 
    52     "count" : 4.0 
     76    "_id" : "gb", 
     77    "count" : 3.0 
    5378} 
    5479 
    5580/* 10 */ 
    5681{ 
    57     "_id" : "gb", 
     82    "_id" : "ca", 
    5883    "count" : 3.0 
    5984} 
     
    6186/* 11 */ 
    6287{ 
    63     "_id" : "ie", 
     88    "_id" : "ua", 
    6489    "count" : 2.0 
    6590} 
     
    6792/* 12 */ 
    6893{ 
    69     "_id" : "ua", 
     94    "_id" : "ie", 
    7095    "count" : 2.0 
    7196} 
     
    80105{ 
    81106    "_id" : "sg", 
     107    "count" : 2.0 
     108} 
     109 
     110/* 15 */ 
     111{ 
     112    "_id" : "unknown", 
     113    "count" : 2.0 
     114} 
     115 
     116/* 16 */ 
     117{ 
     118    "_id" : "gr", 
    82119    "count" : 1.0 
    83120} 
    84121 
    85 /* 15 */ 
     122/* 17 */ 
     123{ 
     124    "_id" : "hk", 
     125    "count" : 1.0 
     126} 
     127 
     128/* 18 */ 
     129{ 
     130    "_id" : "jp", 
     131    "count" : 1.0 
     132} 
     133 
     134/* 19 */ 
     135{ 
     136    "_id" : "bg", 
     137    "count" : 1.0 
     138} 
     139 
     140/* 20 */ 
    86141{ 
    87142    "_id" : "mx", 
     
    89144} 
    90145 
    91 /* 16 */ 
     146/* 21 */ 
    92147{ 
    93     "_id" : "hk", 
     148    "_id" : "ro", 
    94149    "count" : 1.0 
    95150} 
    96  
    97 /* 17 */ 
    98 { 
    99     "_id" : "jp", 
    100     "count" : 1.0 
    101 } 
    102  
    103 /* 18 */ 
    104 { 
    105     "_id" : "se", 
    106     "count" : 1.0 
    107 } 
    108  
    109 /* 19 */ 
    110 { 
    111     "_id" : "gr", 
    112     "count" : 1.0 
    113 } 
    114  
    115 /* 20 */ 
    116 { 
    117     "_id" : "unknown", 
    118     "count" : 1.0 
    119 } 
    120  
    121 /* 21 */ 
    122 { 
    123     "_id" : "bg", 
    124     "count" : 1.0 
    125 } 
  • other-projects/maori-lang-detection/mongodb-data/counts_tentativeNonProductSites.json

    r33806 r33813  
     1/* 
     2 
     3The websites that have some MRI detected AND which are either in NZ or with NZ TLD 
     4or (so if they're from overseas) don't contain /mi or mi.* in URL path. 
     5We'll include Australia, to get the valid "kiwiproperty.com" website, 
     6otherwise the sole exception, included in the result list. 
     7 
     8 
     9db.getCollection('Websites').find({$and: [ 
     10                {numPagesContainingMRI: {$gt: 0}}, 
     11                {$or: [{geoLocationCountryCode: /(NZ|AU)/}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]} 
     12            ]}).count() 
     13 
     14397 
     15 
     16Aggregate results by a count of country codes: 
     17 
     18db.Websites.aggregate([ 
     19    { 
     20        $match: { 
     21            $and: [ 
     22                {numPagesContainingMRI: {$gt: 0}}, 
     23                {$or: [{geoLocationCountryCode: /(NZ|AU)/}, {domain: /\.nz$/}, {urlContainsLangCodeInPath: false}]} 
     24            ] 
     25        } 
     26    }, 
     27    { $unwind: "$geoLocationCountryCode" }, 
     28    { 
     29        $group: { 
     30            _id: {$toLower: '$geoLocationCountryCode'}, 
     31            count: { $sum: 1 } 
     32        } 
     33    }, 
     34    { $sort : { count : -1} } 
     35]); 
     36 
     37*/ 
     38 
    139/* 1 */ 
    240{ 
    341    "_id" : "us", 
    4     "count" : 475.0 
     42    "count" : 181.0 
    543} 
    644 
    745/* 2 */ 
    846{ 
    9     "_id" : "cn", 
    10     "count" : 114.0 
     47    "_id" : "nz", 
     48    "count" : 89.0 
    1149} 
    1250 
    1351/* 3 */ 
    1452{ 
    15     "_id" : "nz", 
    16     "count" : 98.0 
     53    "_id" : "au", 
     54    "count" : 21.0 
    1755} 
    1856 
    1957/* 4 */ 
    2058{ 
     59    "_id" : "de", 
     60    "count" : 19.0 
     61} 
     62 
     63/* 5 */ 
     64{ 
    2165    "_id" : "fr", 
    22     "count" : 36.0 
    23 } 
    24  
    25 /* 5 */ 
    26 { 
    27     "_id" : "de", 
    28     "count" : 26.0 
     66    "count" : 17.0 
    2967} 
    3068 
     
    3270{ 
    3371    "_id" : "nl", 
    34     "count" : 22.0 
     72    "count" : 16.0 
    3573} 
    3674 
    3775/* 7 */ 
    38 { 
    39     "_id" : "au", 
    40     "count" : 17.0 
    41 } 
    42  
    43 /* 8 */ 
    44 { 
    45     "_id" : "ca", 
    46     "count" : 13.0 
    47 } 
    48  
    49 /* 9 */ 
    5076{ 
    5177    "_id" : "dk", 
     
    5379} 
    5480 
     81/* 8 */ 
     82{ 
     83    "_id" : "ca", 
     84    "count" : 7.0 
     85} 
     86 
     87/* 9 */ 
     88{ 
     89    "_id" : "es", 
     90    "count" : 6.0 
     91} 
     92 
    5593/* 10 */ 
    5694{ 
    57     "_id" : "es", 
    58     "count" : 7.0 
     95    "_id" : "gb", 
     96    "count" : 5.0 
    5997} 
    6098 
    6199/* 11 */ 
    62 { 
    63     "_id" : "gb", 
    64     "count" : 7.0 
    65 } 
    66  
    67 /* 12 */ 
    68100{ 
    69101    "_id" : "cz", 
     
    71103} 
    72104 
     105/* 12 */ 
     106{ 
     107    "_id" : "at", 
     108    "count" : 3.0 
     109} 
     110 
    73111/* 13 */ 
    74112{ 
     
    79117/* 14 */ 
    80118{ 
    81     "_id" : "at", 
     119    "_id" : "ro", 
    82120    "count" : 3.0 
    83121} 
    84122 
    85123/* 15 */ 
     124{ 
     125    "_id" : "il", 
     126    "count" : 2.0 
     127} 
     128 
     129/* 16 */ 
    86130{ 
    87131    "_id" : "ch", 
     
    89133} 
    90134 
    91 /* 16 */ 
    92 { 
    93     "_id" : "ro", 
    94     "count" : 2.0 
    95 } 
    96  
    97135/* 17 */ 
    98136{ 
    99     "_id" : "il", 
    100     "count" : 2.0 
     137    "_id" : "bg", 
     138    "count" : 1.0 
    101139} 
    102140 
    103141/* 18 */ 
    104142{ 
     143    "_id" : "sg", 
     144    "count" : 1.0 
     145} 
     146 
     147/* 19 */ 
     148{ 
     149    "_id" : "mx", 
     150    "count" : 1.0 
     151} 
     152 
     153/* 20 */ 
     154{ 
     155    "_id" : "ir", 
     156    "count" : 1.0 
     157} 
     158 
     159/* 21 */ 
     160{ 
     161    "_id" : "cn", 
     162    "count" : 1.0 
     163} 
     164 
     165/* 22 */ 
     166{ 
     167    "_id" : "ie", 
     168    "count" : 1.0 
     169} 
     170 
     171/* 23 */ 
     172{ 
     173    "_id" : "jp", 
     174    "count" : 1.0 
     175} 
     176 
     177/* 24 */ 
     178{ 
     179    "_id" : "fi", 
     180    "count" : 1.0 
     181} 
     182 
     183/* 25 */ 
     184{ 
     185    "_id" : "gr", 
     186    "count" : 1.0 
     187} 
     188 
     189/* 26 */ 
     190{ 
     191    "_id" : "ru", 
     192    "count" : 1.0 
     193} 
     194 
     195/* 27 */ 
     196{ 
    105197    "_id" : "unknown", 
    106     "count" : 2.0 
    107 } 
    108  
    109 /* 19 */ 
    110 { 
    111     "_id" : "hk", 
    112     "count" : 2.0 
    113 } 
    114  
    115 /* 20 */ 
    116 { 
    117     "_id" : "jp", 
    118     "count" : 2.0 
    119 } 
    120  
    121 /* 21 */ 
    122 { 
    123     "_id" : "ie", 
    124     "count" : 2.0 
    125 } 
    126  
    127 /* 22 */ 
    128 { 
    129     "_id" : "ua", 
    130     "count" : 2.0 
    131 } 
    132  
    133 /* 23 */ 
    134 { 
    135     "_id" : "se", 
    136     "count" : 1.0 
    137 } 
    138  
    139 /* 24 */ 
    140 { 
    141     "_id" : "gr", 
    142     "count" : 1.0 
    143 } 
    144  
    145 /* 25 */ 
    146 { 
    147     "_id" : "ru", 
    148     "count" : 1.0 
    149 } 
    150  
    151 /* 26 */ 
    152 { 
    153     "_id" : "eu", 
    154     "count" : 1.0 
    155 } 
    156  
    157 /* 27 */ 
    158 { 
    159     "_id" : "bg", 
    160     "count" : 1.0 
    161 } 
    162  
    163 /* 28 */ 
    164 { 
    165     "_id" : "fi", 
    166     "count" : 1.0 
    167 } 
    168  
    169 /* 29 */ 
    170 { 
    171     "_id" : "sg", 
    172     "count" : 1.0 
    173 } 
    174  
    175 /* 30 */ 
    176 { 
    177     "_id" : "tr", 
    178     "count" : 1.0 
    179 } 
    180  
    181 /* 31 */ 
    182 { 
    183     "_id" : "mx", 
    184     "count" : 1.0 
    185 } 
    186  
    187 /* 32 */ 
    188 { 
    189     "_id" : "ir", 
    190     "count" : 1.0 
    191 } 
     198    "count" : 1.0 
     199} 
     200 
  • other-projects/maori-lang-detection/mongodb-data/geojson-features_allCrawledSites.json

    r33802 r33813  
    1 {"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-127.712891,37.09024],[-127.712891,79.96524],[-63.712891,79.96524],[-63.712891,37.09024],[-127.712891,37.09024]]]},"properties":{"code":"US","count":686,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[53.0,-70.0],[53.0,15.5],[61.0,15.5],[61.0,-70.0],[53.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":171,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,87.09944300000001],[176.885971,87.09944300000001],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":128,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[96.195397,35.86166],[96.195397,66.86166],[112.195397,66.86166],[112.195397,35.86166],[96.195397,35.86166]]]},"properties":{"code":"CN","count":124,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-1.786251,46.227638],[-1.786251,80.727638],[6.213749,80.727638],[6.213749,46.227638],[-1.786251,46.227638]]]},"properties":{"code":"FR","count":69,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.451525999999999,51.165691],[6.451525999999999,77.16569100000001],[14.451526,77.16569100000001],[14.451526,51.165691],[6.451525999999999,51.165691]]]},"properties":{"code":"DE","count":52,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,17.725602],[135.775136,17.725602],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":43,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,83.132633],[7.291266,83.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":31,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,77.13036600000001],[-104.346771,77.13036600000001],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":21,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,73.378051],[-1.4359730000000002,73.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":18,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,46.204824],[140.252924,46.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":10,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,66.26392],[11.501785,66.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":10,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,48.463667],[-1.7492200000000002,48.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":8,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[13.472962,49.817492],[13.472962,56.817492],[17.472962000000003,56.817492],[17.472962000000003,49.817492],[13.472962,49.817492]]]},"properties":{"code":"CZ","count":7,"region":"Czech Republic"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[103.318756,61.52401],[103.318756,68.52401],[107.318756,68.52401],[107.318756,61.52401],[103.318756,61.52401]]]},"properties":{"code":"RU","count":7,"region":"Russia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,29.396428],[116.109497,29.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":7,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[10.56738,41.87194],[10.56738,46.87194],[14.56738,46.87194],[14.56738,41.87194],[10.56738,41.87194]]]},"properties":{"code":"IT","count":5,"region":"Italy"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[16.643501,60.128161],[16.643501,65.128161],[20.643501,65.128161],[20.643501,60.128161],[16.643501,60.128161]]]},"properties":{"code":"SE","count":5,"region":"Sweden"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,53.379433],[33.16558,53.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":5,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,57.41291],[-6.24389,57.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":4,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[76.96288,20.593684],[76.96288,23.593684],[80.96288,23.593684],[80.96288,20.593684],[76.96288,20.593684]]]},"properties":{"code":"IN","count":3,"region":"India"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.227512000000001,46.818188],[6.227512000000001,49.818188],[10.227512,49.818188],[10.227512,46.818188],[6.227512000000001,46.818188]]]},"properties":{"code":"CH","count":3,"region":"Switzerland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[32.851612,31.046051],[32.851612,34.046051],[36.851612,34.046051],[36.851612,31.046051],[32.851612,31.046051]]]},"properties":{"code":"IL","count":3,"region":"Israel"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[12.550072,47.516231],[12.550072,50.516231],[16.550072,50.516231],[16.550072,47.516231],[12.550072,47.516231]]]},"properties":{"code":"AT","count":3,"region":"Austria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,47.943161],[26.96676,47.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":2,"region":"Romania"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-161.777671,-21.236736],[-161.777671,-19.236736],[-157.777671,-19.236736],[-157.777671,-21.236736],[-161.777671,-21.236736]]]},"properties":{"code":"CK","count":2,"region":"Cook Islands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-66.639968,18.420695],[-66.639968,20.420695],[-62.639967999999996,20.420695],[-62.639967999999996,18.420695],[-66.639968,18.420695]]]},"properties":{"code":"VG","count":2,"region":"British Virgin Islands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[20.937506,-30.559482],[20.937506,-28.559482],[24.937506,-28.559482],[24.937506,-30.559482],[20.937506,-30.559482]]]},"properties":{"code":"ZA","count":2,"region":"South Africa"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[17.145136,51.919438],[17.145136,53.919438],[21.145136,53.919438],[21.145136,51.919438],[17.145136,51.919438]]]},"properties":{"code":"PL","count":2,"region":"Poland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.748151,61.92411],[23.748151,62.92411],[27.748151,62.92411],[27.748151,61.92411],[23.748151,61.92411]]]},"properties":{"code":"FI","count":1,"region":"Finland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[17.37439,42.708678],[17.37439,43.708678],[21.37439,43.708678],[21.37439,42.708678],[17.37439,42.708678]]]},"properties":{"code":"ME","count":1,"region":"Montenegro"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.224454,39.399872],[-10.224454,40.399872],[-6.224454,40.399872],[-6.224454,39.399872],[-10.224454,39.399872]]]},"properties":{"code":"PT","count":1,"region":"Portugal"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[51.688046,32.427908],[51.688046,33.427908],[55.688046,33.427908],[55.688046,32.427908],[51.688046,32.427908]]]},"properties":{"code":"IR","count":1,"region":"Iran"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-22.0,50.0],[-22.0,51.0],[-18.0,51.0],[-18.0,50.0],[-22.0,50.0]]]},"properties":{"code":"EU","count":1,"region":"Europe"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[33.243322,38.963745],[33.243322,39.963745],[37.243322,39.963745],[37.243322,38.963745],[33.243322,38.963745]]]},"properties":{"code":"TR","count":1,"region":"Turkey"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[69.876519,-6.343194],[69.876519,-5.343194],[73.876519,-5.343194],[73.876519,-6.343194],[69.876519,-6.343194]]]},"properties":{"code":"IO","count":1,"region":"British Indian Ocean Territory"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-151.406843,-17.679742],[-151.406843,-16.679742],[-147.406843,-16.679742],[-147.406843,-17.679742],[-151.406843,-17.679742]]]},"properties":{"code":"PF","count":1,"region":"French Polynesia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,2.352083],[105.819836,2.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":1,"region":"Singapore"}}]} 
     1{"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-127.712891,37.09024],[-127.712891,80.59024],[-63.712891,80.59024],[-63.712891,37.09024],[-127.712891,37.09024]]]},"properties":{"code":"US","count":696,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[53.0,-70.0],[53.0,16.5],[61.0,16.5],[61.0,-70.0],[53.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":173,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[96.195397,35.86166],[96.195397,67.11166],[112.195397,67.11166],[112.195397,35.86166],[96.195397,35.86166]]]},"properties":{"code":"CN","count":125,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,74.09944300000001],[176.885971,74.09944300000001],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":115,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-1.786251,46.227638],[-1.786251,80.727638],[6.213749,80.727638],[6.213749,46.227638],[-1.786251,46.227638]]]},"properties":{"code":"FR","count":69,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.451525999999999,51.165691],[6.451525999999999,77.16569100000001],[14.451526,77.16569100000001],[14.451526,51.165691],[6.451525999999999,51.165691]]]},"properties":{"code":"DE","count":52,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,17.725602],[135.775136,17.725602],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":43,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,84.132633],[7.291266,84.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":32,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,75.13036600000001],[-104.346771,75.13036600000001],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":19,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,73.378051],[-1.4359730000000002,73.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":18,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,66.26392],[11.501785,66.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":10,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,46.204824],[140.252924,46.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":10,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,48.463667],[-1.7492200000000002,48.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":8,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[103.318756,61.52401],[103.318756,68.52401],[107.318756,68.52401],[107.318756,61.52401],[103.318756,61.52401]]]},"properties":{"code":"RU","count":7,"region":"Russia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,29.396428],[116.109497,29.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":7,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[13.472962,49.817492],[13.472962,56.817492],[17.472962000000003,56.817492],[17.472962000000003,49.817492],[13.472962,49.817492]]]},"properties":{"code":"CZ","count":7,"region":"Czech Republic"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,53.379433],[33.16558,53.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":5,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,57.41291],[-6.24389,57.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":4,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[16.643501,60.128161],[16.643501,64.128161],[20.643501,64.128161],[20.643501,60.128161],[16.643501,60.128161]]]},"properties":{"code":"SE","count":4,"region":"Sweden"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[10.56738,41.87194],[10.56738,45.87194],[14.56738,45.87194],[14.56738,41.87194],[10.56738,41.87194]]]},"properties":{"code":"IT","count":4,"region":"Italy"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,48.943161],[26.96676,48.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":3,"region":"Romania"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,4.352083],[105.819836,4.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":3,"region":"Singapore"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[12.550072,47.516231],[12.550072,50.516231],[16.550072,50.516231],[16.550072,47.516231],[12.550072,47.516231]]]},"properties":{"code":"AT","count":3,"region":"Austria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.227512000000001,46.818188],[6.227512000000001,49.818188],[10.227512,49.818188],[10.227512,46.818188],[6.227512000000001,46.818188]]]},"properties":{"code":"CH","count":3,"region":"Switzerland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[32.851612,31.046051],[32.851612,34.046051],[36.851612,34.046051],[36.851612,31.046051],[32.851612,31.046051]]]},"properties":{"code":"IL","count":3,"region":"Israel"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[76.96288,20.593684],[76.96288,23.593684],[80.96288,23.593684],[80.96288,20.593684],[76.96288,20.593684]]]},"properties":{"code":"IN","count":3,"region":"India"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[17.145136,51.919438],[17.145136,53.919438],[21.145136,53.919438],[21.145136,51.919438],[17.145136,51.919438]]]},"properties":{"code":"PL","count":2,"region":"Poland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[20.937506,-30.559482],[20.937506,-28.559482],[24.937506,-28.559482],[24.937506,-30.559482],[20.937506,-30.559482]]]},"properties":{"code":"ZA","count":2,"region":"South Africa"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-66.639968,18.420695],[-66.639968,20.420695],[-62.639967999999996,20.420695],[-62.639967999999996,18.420695],[-66.639968,18.420695]]]},"properties":{"code":"VG","count":2,"region":"British Virgin Islands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-161.777671,-21.236736],[-161.777671,-19.236736],[-157.777671,-19.236736],[-157.777671,-21.236736],[-161.777671,-21.236736]]]},"properties":{"code":"CK","count":2,"region":"Cook Islands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-151.406843,-17.679742],[-151.406843,-16.679742],[-147.406843,-16.679742],[-147.406843,-17.679742],[-151.406843,-17.679742]]]},"properties":{"code":"PF","count":1,"region":"French Polynesia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[69.876519,-6.343194],[69.876519,-5.343194],[73.876519,-5.343194],[73.876519,-6.343194],[69.876519,-6.343194]]]},"properties":{"code":"IO","count":1,"region":"British Indian Ocean Territory"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[33.243322,38.963745],[33.243322,39.963745],[37.243322,39.963745],[37.243322,38.963745],[33.243322,38.963745]]]},"properties":{"code":"TR","count":1,"region":"Turkey"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[17.37439,42.708678],[17.37439,43.708678],[21.37439,43.708678],[21.37439,42.708678],[17.37439,42.708678]]]},"properties":{"code":"ME","count":1,"region":"Montenegro"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.748151,61.92411],[23.748151,62.92411],[27.748151,62.92411],[27.748151,61.92411],[23.748151,61.92411]]]},"properties":{"code":"FI","count":1,"region":"Finland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-22.0,50.0],[-22.0,51.0],[-18.0,51.0],[-18.0,50.0],[-22.0,50.0]]]},"properties":{"code":"EU","count":1,"region":"Europe"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[51.688046,32.427908],[51.688046,33.427908],[55.688046,33.427908],[55.688046,32.427908],[51.688046,32.427908]]]},"properties":{"code":"IR","count":1,"region":"Iran"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.224454,39.399872],[-10.224454,40.399872],[-6.224454,40.399872],[-6.224454,39.399872],[-10.224454,39.399872]]]},"properties":{"code":"PT","count":1,"region":"Portugal"}}]} 
  • other-projects/maori-lang-detection/mongodb-data/geojson-features_sitesWithPagesContainingMRI.json

    r33803 r33813  
    1 {"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-127.712891,37.09024],[-127.712891,67.02774],[-63.712891,67.02774],[-63.712891,37.09024],[-127.712891,37.09024]]]},"properties":{"code":"US","count":479,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[96.195397,35.86166],[96.195397,64.36166],[112.195397,64.36166],[112.195397,35.86166],[96.195397,35.86166]]]},"properties":{"code":"CN","count":114,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,57.099443],[176.885971,57.099443],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":98,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[0.21374899999999997,46.227638],[0.21374899999999997,82.227638],[4.213749,82.227638],[4.213749,46.227638],[0.21374899999999997,46.227638]]]},"properties":{"code":"FR","count":36,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[8.451526,51.165691],[8.451526,78.16569100000001],[12.451526,78.16569100000001],[12.451526,51.165691],[8.451526,51.165691]]]},"properties":{"code":"DE","count":27,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,74.132633],[7.291266,74.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":22,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,-5.2743980000000015],[135.775136,-5.2743980000000015],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":20,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,69.13036600000001],[-104.346771,69.13036600000001],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":13,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,64.26392],[11.501785,64.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":8,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,62.378051],[-1.4359730000000002,62.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":7,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,47.463667],[-1.7492200000000002,47.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":7,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[13.472962,49.817492],[13.472962,53.817492],[17.472962000000003,53.817492],[17.472962000000003,49.817492],[13.472962,49.817492]]]},"properties":{"code":"CZ","count":4,"region":"Czech Republic"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[10.56738,41.87194],[10.56738,44.87194],[14.56738,44.87194],[14.56738,41.87194],[10.56738,41.87194]]]},"properties":{"code":"IT","count":3,"region":"Italy"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[12.550072,47.516231],[12.550072,50.516231],[16.550072,50.516231],[16.550072,47.516231],[12.550072,47.516231]]]},"properties":{"code":"AT","count":3,"region":"Austria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[32.851612,31.046051],[32.851612,33.046051],[36.851612,33.046051],[36.851612,31.046051],[32.851612,31.046051]]]},"properties":{"code":"IL","count":2,"region":"Israel"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.227512000000001,46.818188],[6.227512000000001,48.818188],[10.227512,48.818188],[10.227512,46.818188],[6.227512000000001,46.818188]]]},"properties":{"code":"CH","count":2,"region":"Switzerland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,47.943161],[26.96676,47.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":2,"region":"Romania"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[103.318756,61.52401],[103.318756,63.52401],[107.318756,63.52401],[107.318756,61.52401],[103.318756,61.52401]]]},"properties":{"code":"RU","count":2,"region":"Russia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,55.41291],[-6.24389,55.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":2,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,38.204824],[140.252924,38.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":2,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,24.396428],[116.109497,24.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":2,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,50.379433],[33.16558,50.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":2,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[55.0,-70.0],[55.0,-68.0],[59.0,-68.0],[59.0,-70.0],[55.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":2,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.748151,61.92411],[23.748151,62.92411],[27.748151,62.92411],[27.748151,61.92411],[23.748151,61.92411]]]},"properties":{"code":"FI","count":1,"region":"Finland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[16.643501,60.128161],[16.643501,61.128161],[20.643501,61.128161],[20.643501,60.128161],[16.643501,60.128161]]]},"properties":{"code":"SE","count":1,"region":"Sweden"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-22.0,50.0],[-22.0,51.0],[-18.0,51.0],[-18.0,50.0],[-22.0,50.0]]]},"properties":{"code":"EU","count":1,"region":"Europe"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[51.688046,32.427908],[51.688046,33.427908],[55.688046,33.427908],[55.688046,32.427908],[51.688046,32.427908]]]},"properties":{"code":"IR","count":1,"region":"Iran"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[33.243322,38.963745],[33.243322,39.963745],[37.243322,39.963745],[37.243322,38.963745],[33.243322,38.963745]]]},"properties":{"code":"TR","count":1,"region":"Turkey"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,2.352083],[105.819836,2.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":1,"region":"Singapore"}}]} 
     1{"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-127.712891,37.09024],[-127.712891,67.46524],[-63.712891,67.46524],[-63.712891,37.09024],[-127.712891,37.09024]]]},"properties":{"code":"US","count":486,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[96.195397,35.86166],[96.195397,64.36166],[112.195397,64.36166],[112.195397,35.86166],[96.195397,35.86166]]]},"properties":{"code":"CN","count":114,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,48.099443],[176.885971,48.099443],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":89,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[0.21374899999999997,46.227638],[0.21374899999999997,82.227638],[4.213749,82.227638],[4.213749,46.227638],[0.21374899999999997,46.227638]]]},"properties":{"code":"FR","count":36,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[8.451526,51.165691],[8.451526,78.16569100000001],[12.451526,78.16569100000001],[12.451526,51.165691],[8.451526,51.165691]]]},"properties":{"code":"DE","count":27,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,74.132633],[7.291266,74.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":22,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,-4.2743980000000015],[135.775136,-4.2743980000000015],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":21,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,68.13036600000001],[-104.346771,68.13036600000001],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":12,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,64.26392],[11.501785,64.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":8,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,47.463667],[-1.7492200000000002,47.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":7,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,62.378051],[-1.4359730000000002,62.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":7,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[13.472962,49.817492],[13.472962,53.817492],[17.472962000000003,53.817492],[17.472962000000003,49.817492],[13.472962,49.817492]]]},"properties":{"code":"CZ","count":4,"region":"Czech Republic"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[55.0,-70.0],[55.0,-67.0],[59.0,-67.0],[59.0,-70.0],[55.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":3,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[12.550072,47.516231],[12.550072,50.516231],[16.550072,50.516231],[16.550072,47.516231],[12.550072,47.516231]]]},"properties":{"code":"AT","count":3,"region":"Austria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,48.943161],[26.96676,48.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":3,"region":"Romania"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[10.56738,41.87194],[10.56738,44.87194],[14.56738,44.87194],[14.56738,41.87194],[10.56738,41.87194]]]},"properties":{"code":"IT","count":3,"region":"Italy"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,3.352083],[105.819836,3.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":2,"region":"Singapore"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,38.204824],[140.252924,38.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":2,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,55.41291],[-6.24389,55.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":2,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,24.396428],[116.109497,24.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":2,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,50.379433],[33.16558,50.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":2,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[103.318756,61.52401],[103.318756,63.52401],[107.318756,63.52401],[107.318756,61.52401],[103.318756,61.52401]]]},"properties":{"code":"RU","count":2,"region":"Russia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.227512000000001,46.818188],[6.227512000000001,48.818188],[10.227512,48.818188],[10.227512,46.818188],[6.227512000000001,46.818188]]]},"properties":{"code":"CH","count":2,"region":"Switzerland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[32.851612,31.046051],[32.851612,33.046051],[36.851612,33.046051],[36.851612,31.046051],[32.851612,31.046051]]]},"properties":{"code":"IL","count":2,"region":"Israel"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[33.243322,38.963745],[33.243322,39.963745],[37.243322,39.963745],[37.243322,38.963745],[33.243322,38.963745]]]},"properties":{"code":"TR","count":1,"region":"Turkey"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[51.688046,32.427908],[51.688046,33.427908],[55.688046,33.427908],[55.688046,32.427908],[51.688046,32.427908]]]},"properties":{"code":"IR","count":1,"region":"Iran"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-22.0,50.0],[-22.0,51.0],[-18.0,51.0],[-18.0,50.0],[-22.0,50.0]]]},"properties":{"code":"EU","count":1,"region":"Europe"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.748151,61.92411],[23.748151,62.92411],[27.748151,62.92411],[27.748151,61.92411],[23.748151,61.92411]]]},"properties":{"code":"FI","count":1,"region":"Finland"}}]} 
  • other-projects/maori-lang-detection/mongodb-data/geojson-features_sitesWithPagesInMRI.json

    r33805 r33813  
    1 {"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-103.712891,37.09024],[-103.712891,88.84024],[-87.712891,88.84024],[-87.712891,37.09024],[-103.712891,37.09024]]]},"properties":{"code":"US","count":207,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,12.099443],[176.885971,12.099443],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":53,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[102.195397,35.86166],[102.195397,67.86166],[106.195397,67.86166],[106.195397,35.86166],[102.195397,35.86166]]]},"properties":{"code":"CN","count":32,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[0.21374899999999997,46.227638],[0.21374899999999997,64.227638],[4.213749,64.227638],[4.213749,46.227638],[0.21374899999999997,46.227638]]]},"properties":{"code":"FR","count":18,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,-14.274398000000001],[135.775136,-14.274398000000001],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":11,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,62.132633],[7.291266,62.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":10,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[8.451526,51.165691],[8.451526,56.165691],[12.451526,56.165691],[12.451526,51.165691],[8.451526,51.165691]]]},"properties":{"code":"DE","count":5,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,60.26392],[11.501785,60.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":4,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,60.130366],[-104.346771,60.130366],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":4,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,58.378051],[-1.4359730000000002,58.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":3,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,55.41291],[-6.24389,55.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":2,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,50.379433],[33.16558,50.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":2,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,42.463667],[-1.7492200000000002,42.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":2,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,2.352083],[105.819836,2.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":1,"region":"Singapore"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,23.396428],[116.109497,23.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":1,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,37.204824],[140.252924,37.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":1,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[16.643501,60.128161],[16.643501,61.128161],[20.643501,61.128161],[20.643501,60.128161],[16.643501,60.128161]]]},"properties":{"code":"SE","count":1,"region":"Sweden"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[55.0,-70.0],[55.0,-69.0],[59.0,-69.0],[59.0,-70.0],[55.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":1,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}}]} 
     1{"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-103.712891,37.09024],[-103.712891,88.59024],[-87.712891,88.59024],[-87.712891,37.09024],[-103.712891,37.09024]]]},"properties":{"code":"US","count":206,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,12.099443],[176.885971,12.099443],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":53,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[102.195397,35.86166],[102.195397,67.86166],[106.195397,67.86166],[106.195397,35.86166],[102.195397,35.86166]]]},"properties":{"code":"CN","count":32,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[0.21374899999999997,46.227638],[0.21374899999999997,64.227638],[4.213749,64.227638],[4.213749,46.227638],[0.21374899999999997,46.227638]]]},"properties":{"code":"FR","count":18,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,-14.274398000000001],[135.775136,-14.274398000000001],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":11,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,62.132633],[7.291266,62.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":10,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[8.451526,51.165691],[8.451526,56.165691],[12.451526,56.165691],[12.451526,51.165691],[8.451526,51.165691]]]},"properties":{"code":"DE","count":5,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,60.26392],[11.501785,60.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":4,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,58.378051],[-1.4359730000000002,58.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":3,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,59.130366],[-104.346771,59.130366],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":3,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,50.379433],[33.16558,50.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":2,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,55.41291],[-6.24389,55.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":2,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,42.463667],[-1.7492200000000002,42.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":2,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,3.352083],[105.819836,3.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":2,"region":"Singapore"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[55.0,-70.0],[55.0,-68.0],[59.0,-68.0],[59.0,-70.0],[55.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":2,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,23.396428],[116.109497,23.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":1,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,37.204824],[140.252924,37.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":1,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,46.943161],[26.96676,46.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":1,"region":"Romania"}}]} 
  • other-projects/maori-lang-detection/mongodb-data/geojson-features_tentativeNonProductSites.json

    r33806 r33813  
    1 {"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-127.712891,37.09024],[-127.712891,66.77774],[-63.712891,66.77774],[-63.712891,37.09024],[-127.712891,37.09024]]]},"properties":{"code":"US","count":475,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[96.195397,35.86166],[96.195397,64.36166],[112.195397,64.36166],[112.195397,35.86166],[96.195397,35.86166]]]},"properties":{"code":"CN","count":114,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,57.099443],[176.885971,57.099443],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":98,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[0.21374899999999997,46.227638],[0.21374899999999997,82.227638],[4.213749,82.227638],[4.213749,46.227638],[0.21374899999999997,46.227638]]]},"properties":{"code":"FR","count":36,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[8.451526,51.165691],[8.451526,77.16569100000001],[12.451526,77.16569100000001],[12.451526,51.165691],[8.451526,51.165691]]]},"properties":{"code":"DE","count":26,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,74.132633],[7.291266,74.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":22,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,-8.274398000000001],[135.775136,-8.274398000000001],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":17,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,69.13036600000001],[-104.346771,69.13036600000001],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":13,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,64.26392],[11.501785,64.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":8,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,47.463667],[-1.7492200000000002,47.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":7,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,62.378051],[-1.4359730000000002,62.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":7,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[13.472962,49.817492],[13.472962,53.817492],[17.472962000000003,53.817492],[17.472962000000003,49.817492],[13.472962,49.817492]]]},"properties":{"code":"CZ","count":4,"region":"Czech Republic"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[10.56738,41.87194],[10.56738,44.87194],[14.56738,44.87194],[14.56738,41.87194],[10.56738,41.87194]]]},"properties":{"code":"IT","count":3,"region":"Italy"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[12.550072,47.516231],[12.550072,50.516231],[16.550072,50.516231],[16.550072,47.516231],[12.550072,47.516231]]]},"properties":{"code":"AT","count":3,"region":"Austria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.227512000000001,46.818188],[6.227512000000001,48.818188],[10.227512,48.818188],[10.227512,46.818188],[6.227512000000001,46.818188]]]},"properties":{"code":"CH","count":2,"region":"Switzerland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,47.943161],[26.96676,47.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":2,"region":"Romania"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[32.851612,31.046051],[32.851612,33.046051],[36.851612,33.046051],[36.851612,31.046051],[32.851612,31.046051]]]},"properties":{"code":"IL","count":2,"region":"Israel"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[55.0,-70.0],[55.0,-68.0],[59.0,-68.0],[59.0,-70.0],[55.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":2,"region":"UNKNOWN"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[112.109497,22.396428],[112.109497,24.396428],[116.109497,24.396428],[116.109497,22.396428],[112.109497,22.396428]]]},"properties":{"code":"HK","count":2,"region":"Hong Kong"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,38.204824],[140.252924,38.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":2,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,55.41291],[-6.24389,55.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":2,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[29.16558,48.379433],[29.16558,50.379433],[33.16558,50.379433],[33.16558,48.379433],[29.16558,48.379433]]]},"properties":{"code":"UA","count":2,"region":"Ukraine"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[16.643501,60.128161],[16.643501,61.128161],[20.643501,61.128161],[20.643501,60.128161],[16.643501,60.128161]]]},"properties":{"code":"SE","count":1,"region":"Sweden"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[103.318756,61.52401],[103.318756,62.52401],[107.318756,62.52401],[107.318756,61.52401],[103.318756,61.52401]]]},"properties":{"code":"RU","count":1,"region":"Russia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-22.0,50.0],[-22.0,51.0],[-18.0,51.0],[-18.0,50.0],[-22.0,50.0]]]},"properties":{"code":"EU","count":1,"region":"Europe"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.748151,61.92411],[23.748151,62.92411],[27.748151,62.92411],[27.748151,61.92411],[23.748151,61.92411]]]},"properties":{"code":"FI","count":1,"region":"Finland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,2.352083],[105.819836,2.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":1,"region":"Singapore"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[33.243322,38.963745],[33.243322,39.963745],[37.243322,39.963745],[37.243322,38.963745],[33.243322,38.963745]]]},"properties":{"code":"TR","count":1,"region":"Turkey"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[51.688046,32.427908],[51.688046,33.427908],[55.688046,33.427908],[55.688046,32.427908],[51.688046,32.427908]]]},"properties":{"code":"IR","count":1,"region":"Iran"}}]} 
     1{"type":"FeatureCollection","features":[{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-103.712891,37.09024],[-103.712891,82.34024],[-87.712891,82.34024],[-87.712891,37.09024],[-103.712891,37.09024]]]},"properties":{"code":"US","count":181,"region":"United States"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[172.885971,-40.900557],[172.885971,48.099443],[176.885971,48.099443],[176.885971,-40.900557],[172.885971,-40.900557]]]},"properties":{"code":"NZ","count":89,"region":"New Zealand"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[131.775136,-25.274398],[131.775136,-4.2743980000000015],[135.775136,-4.2743980000000015],[135.775136,-25.274398],[131.775136,-25.274398]]]},"properties":{"code":"AU","count":21,"region":"Australia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[8.451526,51.165691],[8.451526,70.16569100000001],[12.451526,70.16569100000001],[12.451526,51.165691],[8.451526,51.165691]]]},"properties":{"code":"DE","count":19,"region":"Germany"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[0.21374899999999997,46.227638],[0.21374899999999997,63.227638],[4.213749,63.227638],[4.213749,46.227638],[0.21374899999999997,46.227638]]]},"properties":{"code":"FR","count":17,"region":"France"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[3.2912660000000002,52.132633],[3.2912660000000002,68.132633],[7.291266,68.132633],[7.291266,52.132633],[3.2912660000000002,52.132633]]]},"properties":{"code":"NL","count":16,"region":"Netherlands"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[7.501785,56.26392],[7.501785,64.26392],[11.501785,64.26392],[11.501785,56.26392],[7.501785,56.26392]]]},"properties":{"code":"DK","count":8,"region":"Denmark"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-108.346771,56.130366],[-108.346771,63.130366],[-104.346771,63.130366],[-104.346771,56.130366],[-108.346771,56.130366]]]},"properties":{"code":"CA","count":7,"region":"Canada"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.74922,40.463667],[-5.74922,46.463667],[-1.7492200000000002,46.463667],[-1.7492200000000002,40.463667],[-5.74922,40.463667]]]},"properties":{"code":"ES","count":6,"region":"Spain"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-5.435973000000001,55.378051],[-5.435973000000001,60.378051],[-1.4359730000000002,60.378051],[-1.4359730000000002,55.378051],[-5.435973000000001,55.378051]]]},"properties":{"code":"GB","count":5,"region":"United Kingdom"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[13.472962,49.817492],[13.472962,53.817492],[17.472962000000003,53.817492],[17.472962000000003,49.817492],[13.472962,49.817492]]]},"properties":{"code":"CZ","count":4,"region":"Czech Republic"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[12.550072,47.516231],[12.550072,50.516231],[16.550072,50.516231],[16.550072,47.516231],[12.550072,47.516231]]]},"properties":{"code":"AT","count":3,"region":"Austria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[10.56738,41.87194],[10.56738,44.87194],[14.56738,44.87194],[14.56738,41.87194],[10.56738,41.87194]]]},"properties":{"code":"IT","count":3,"region":"Italy"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[22.96676,45.943161],[22.96676,48.943161],[26.96676,48.943161],[26.96676,45.943161],[22.96676,45.943161]]]},"properties":{"code":"RO","count":3,"region":"Romania"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[32.851612,31.046051],[32.851612,33.046051],[36.851612,33.046051],[36.851612,31.046051],[32.851612,31.046051]]]},"properties":{"code":"IL","count":2,"region":"Israel"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[6.227512000000001,46.818188],[6.227512000000001,48.818188],[10.227512,48.818188],[10.227512,46.818188],[6.227512000000001,46.818188]]]},"properties":{"code":"CH","count":2,"region":"Switzerland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.48583,42.733883],[23.48583,43.733883],[27.48583,43.733883],[27.48583,42.733883],[23.48583,42.733883]]]},"properties":{"code":"BG","count":1,"region":"Bulgaria"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[101.819836,1.352083],[101.819836,2.352083],[105.819836,2.352083],[105.819836,1.352083],[101.819836,1.352083]]]},"properties":{"code":"SG","count":1,"region":"Singapore"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-104.552784,23.634501],[-104.552784,24.634501],[-100.552784,24.634501],[-100.552784,23.634501],[-104.552784,23.634501]]]},"properties":{"code":"MX","count":1,"region":"Mexico"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[51.688046,32.427908],[51.688046,33.427908],[55.688046,33.427908],[55.688046,32.427908],[51.688046,32.427908]]]},"properties":{"code":"IR","count":1,"region":"Iran"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[102.195397,35.86166],[102.195397,36.86166],[106.195397,36.86166],[106.195397,35.86166],[102.195397,35.86166]]]},"properties":{"code":"CN","count":1,"region":"China"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[-10.24389,53.41291],[-10.24389,54.41291],[-6.24389,54.41291],[-6.24389,53.41291],[-10.24389,53.41291]]]},"properties":{"code":"IE","count":1,"region":"Ireland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[136.252924,36.204824],[136.252924,37.204824],[140.252924,37.204824],[140.252924,36.204824],[136.252924,36.204824]]]},"properties":{"code":"JP","count":1,"region":"Japan"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[23.748151,61.92411],[23.748151,62.92411],[27.748151,62.92411],[27.748151,61.92411],[23.748151,61.92411]]]},"properties":{"code":"FI","count":1,"region":"Finland"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[19.824312,39.074208],[19.824312,40.074208],[23.824312,40.074208],[23.824312,39.074208],[19.824312,39.074208]]]},"properties":{"code":"GR","count":1,"region":"Greece"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[103.318756,61.52401],[103.318756,62.52401],[107.318756,62.52401],[107.318756,61.52401],[103.318756,61.52401]]]},"properties":{"code":"RU","count":1,"region":"Russia"}},{"type":"Feature","geometry":{"type":"Polygon","coordinates":[[[55.0,-70.0],[55.0,-69.0],[59.0,-69.0],[59.0,-70.0],[55.0,-70.0]]]},"properties":{"code":"UNKNOWN","count":1,"region":"UNKNOWN"}}]} 
  • other-projects/maori-lang-detection/mongodb-data/multipoint_allCrawledSites.json

    r33802 r33813  
    33    [-95.712891,37.09024], 
    44    [57.0,-70.0], 
     5    [104.195397,35.86166], 
    56    [174.885971,-40.900557], 
    6     [104.195397,35.86166], 
    77    [2.213749,46.227638], 
    88    [10.451526,51.165691], 
     
    1111    [-106.346771,56.130366], 
    1212    [-3.435973,55.378051], 
     13    [9.501785,56.26392], 
    1314    [138.252924,36.204824], 
    14     [9.501785,56.26392], 
    1515    [-3.74922,40.463667], 
    16     [15.472962,49.817492], 
    1716    [105.318756,61.52401], 
    1817    [114.109497,22.396428], 
    19     [12.56738,41.87194], 
    20     [18.643501,60.128161], 
     18    [15.472962,49.817492], 
    2119    [31.16558,48.379433], 
    2220    [-8.24389,53.41291], 
    23     [78.96288,20.593684], 
     21    [18.643501,60.128161], 
     22    [12.56738,41.87194], 
     23    [24.96676,45.943161], 
     24    [103.819836,1.352083], 
     25    [14.550072,47.516231], 
    2426    [8.227512,46.818188], 
    2527    [34.851612,31.046051], 
    26     [14.550072,47.516231], 
    27     [24.96676,45.943161], 
     28    [78.96288,20.593684], 
     29    [19.145136,51.919438], 
     30    [22.937506,-30.559482], 
     31    [-64.639968,18.420695], 
    2832    [-159.777671,-21.236736], 
    29     [-64.639968,18.420695], 
    30     [22.937506,-30.559482], 
    31     [19.145136,51.919438], 
     33    [25.48583,42.733883], 
     34    [-149.406843,-17.679742], 
     35    [71.876519,-6.343194], 
    3236    [21.824312,39.074208], 
     37    [-102.552784,23.634501], 
     38    [35.243322,38.963745], 
     39    [19.37439,42.708678], 
    3340    [25.748151,61.92411], 
    34     [19.37439,42.708678], 
    35     [-8.224454,39.399872], 
     41    [-20.0,50.0], 
    3642    [53.688046,32.427908], 
    37     [-20.0,50.0], 
    38     [35.243322,38.963745], 
    39     [-102.552784,23.634501], 
    40     [71.876519,-6.343194], 
    41     [-149.406843,-17.679742], 
    42     [25.48583,42.733883], 
    43     [103.819836,1.352083] 
     43    [-8.224454,39.399872] 
    4444]} 
  • other-projects/maori-lang-detection/mongodb-data/multipoint_sitesWithPagesContainingMRI.json

    r33803 r33813  
    1010    [-106.346771,56.130366], 
    1111    [9.501785,56.26392], 
     12    [-3.74922,40.463667], 
    1213    [-3.435973,55.378051], 
    13     [-3.74922,40.463667], 
    1414    [15.472962,49.817492], 
     15    [57.0,-70.0], 
     16    [14.550072,47.516231], 
     17    [24.96676,45.943161], 
    1518    [12.56738,41.87194], 
    16     [14.550072,47.516231], 
    17     [34.851612,31.046051], 
    18     [8.227512,46.818188], 
    19     [24.96676,45.943161], 
    20     [105.318756,61.52401], 
     19    [103.819836,1.352083], 
     20    [138.252924,36.204824], 
    2121    [-8.24389,53.41291], 
    22     [138.252924,36.204824], 
    2322    [114.109497,22.396428], 
    2423    [31.16558,48.379433], 
    25     [57.0,-70.0], 
     24    [105.318756,61.52401], 
     25    [8.227512,46.818188], 
     26    [34.851612,31.046051], 
     27    [35.243322,38.963745], 
    2628    [-102.552784,23.634501], 
    27     [25.748151,61.92411], 
    28     [18.643501,60.128161], 
     29    [53.688046,32.427908], 
    2930    [21.824312,39.074208], 
     31    [25.48583,42.733883], 
    3032    [-20.0,50.0], 
    31     [25.48583,42.733883], 
    32     [53.688046,32.427908], 
    33     [35.243322,38.963745], 
    34     [103.819836,1.352083] 
     33    [25.748151,61.92411] 
    3534]} 
  • other-projects/maori-lang-detection/mongodb-data/multipoint_sitesWithPagesInMRI.json

    r33805 r33813  
    99    [10.451526,51.165691], 
    1010    [9.501785,56.26392], 
     11    [-3.435973,55.378051], 
    1112    [-106.346771,56.130366], 
    12     [-3.435973,55.378051], 
     13    [31.16558,48.379433], 
    1314    [-8.24389,53.41291], 
    14     [31.16558,48.379433], 
    1515    [-3.74922,40.463667], 
    1616    [103.819836,1.352083], 
    17     [-102.552784,23.634501], 
     17    [57.0,-70.0], 
     18    [21.824312,39.074208], 
    1819    [114.109497,22.396428], 
    1920    [138.252924,36.204824], 
    20     [18.643501,60.128161], 
    21     [21.824312,39.074208], 
    22     [57.0,-70.0], 
    23     [25.48583,42.733883] 
     21    [25.48583,42.733883], 
     22    [-102.552784,23.634501], 
     23    [24.96676,45.943161] 
    2424]} 
  • other-projects/maori-lang-detection/mongodb-data/multipoint_tentativeNonProductSites.json

    r33806 r33813  
    22[ 
    33    [-95.712891,37.09024], 
    4     [104.195397,35.86166], 
    54    [174.885971,-40.900557], 
     5    [133.775136,-25.274398], 
     6    [10.451526,51.165691], 
    67    [2.213749,46.227638], 
    7     [10.451526,51.165691], 
    88    [5.291266,52.132633], 
    9     [133.775136,-25.274398], 
     9    [9.501785,56.26392], 
    1010    [-106.346771,56.130366], 
    11     [9.501785,56.26392], 
    1211    [-3.74922,40.463667], 
    1312    [-3.435973,55.378051], 
    1413    [15.472962,49.817492], 
     14    [14.550072,47.516231], 
    1515    [12.56738,41.87194], 
    16     [14.550072,47.516231], 
    17     [8.227512,46.818188], 
    1816    [24.96676,45.943161], 
    1917    [34.851612,31.046051], 
    20     [57.0,-70.0], 
    21     [114.109497,22.396428], 
     18    [8.227512,46.818188], 
     19    [25.48583,42.733883], 
     20    [103.819836,1.352083], 
     21    [-102.552784,23.634501], 
     22    [53.688046,32.427908], 
     23    [104.195397,35.86166], 
     24    [-8.24389,53.41291], 
    2225    [138.252924,36.204824], 
    23     [-8.24389,53.41291], 
    24     [31.16558,48.379433], 
    25     [18.643501,60.128161], 
     26    [25.748151,61.92411], 
    2627    [21.824312,39.074208], 
    2728    [105.318756,61.52401], 
    28     [-20.0,50.0], 
    29     [25.48583,42.733883], 
    30     [25.748151,61.92411], 
    31     [103.819836,1.352083], 
    32     [35.243322,38.963745], 
    33     [-102.552784,23.634501], 
    34     [53.688046,32.427908] 
     29    [57.0,-70.0] 
    3530]}