I use foursquare because it contains a lot of POIs.
Which database do you use when you need a point of interest in North America?
[GIS] POI Datasets for North America
apidatapoint-of-interest
Related Solutions
Try wrapping the features in the following:
{"type":"FeatureCollection","features":[
****ALL THE DATA HERE -- COMMA SEPARATED****
]}
and separating each feature with a comma.
For example, here's the first two features from the IE dataset:
{"geometry": {"type": "Point", "coordinates": [-6.422587, 53.293363]}, "type": "Feature", "id": "SG_41u80gC4971D4Gc0Fv1p8q_53.293363_-6.422587@1308163237", "properties": {"website": "http://www.buongiorno.com", "city": "Dublin", "name": "Buongiorno", "tags": ["mobile", "vas", "community", "social-networking", "connected-devices", "android", "tablets", "smartphones"], "country": "IE", "classifiers": [{"category": "Professional", "type": "Services", "subcategory": "Computer Services"}], "href": "http://api.simplegeo.com/1.0/features/SG_41u80gC4971D4Gc0Fv1p8q_53.293363_-6.422587@1308163237.json", "address": "Lake Drive City West Digital Park", "owner": "simplegeo", "postcode": "3050"}}
{"geometry": {"type": "Point", "coordinates": [-6.250848, 53.339347]}, "type": "Feature", "id": "SG_1dnWbWIg8hX3VyfUKIDRz9_53.339347_-6.250848@1308163237", "properties": {"website": "http://www.simchronise.com", "city": "Dublin", "name": "SIMchronise", "tags": ["mobile-solutions", "mobile-data", "data-synchronisation", "mobile-security", "backup", "restore", "mobile-contacts", "web-2.0-addressbook", "phonebackup"], "country": "IE", "classifiers": [{"category": "Professional", "type": "Services", "subcategory": "Computer Services"}], "href": "http://api.simplegeo.com/1.0/features/SG_1dnWbWIg8hX3VyfUKIDRz9_53.339347_-6.250848@1308163237.json", "address": "78 Merrion Square", "owner": "simplegeo", "postcode": "Dublin2"}}
... and here they are as a GeoJSON FeatureCollection:
{"type":"FeatureCollection","features":[
{"geometry": {"type": "Point", "coordinates": [-6.422587, 53.293363]}, "type": "Feature", "id": "SG_41u80gC4971D4Gc0Fv1p8q_53.293363_-6.422587@1308163237", "properties": {"website": "http://www.buongiorno.com", "city": "Dublin", "name": "Buongiorno", "tags": ["mobile", "vas", "community", "social-networking", "connected-devices", "android", "tablets", "smartphones"], "country": "IE", "classifiers": [{"category": "Professional", "type": "Services", "subcategory": "Computer Services"}], "href": "http://api.simplegeo.com/1.0/features/SG_41u80gC4971D4Gc0Fv1p8q_53.293363_-6.422587@1308163237.json", "address": "Lake Drive City West Digital Park", "owner": "simplegeo", "postcode": "3050"}},
{"geometry": {"type": "Point", "coordinates": [-6.250848, 53.339347]}, "type": "Feature", "id": "SG_1dnWbWIg8hX3VyfUKIDRz9_53.339347_-6.250848@1308163237", "properties": {"website": "http://www.simchronise.com", "city": "Dublin", "name": "SIMchronise", "tags": ["mobile-solutions", "mobile-data", "data-synchronisation", "mobile-security", "backup", "restore", "mobile-contacts", "web-2.0-addressbook", "phonebackup"], "country": "IE", "classifiers": [{"category": "Professional", "type": "Services", "subcategory": "Computer Services"}], "href": "http://api.simplegeo.com/1.0/features/SG_1dnWbWIg8hX3VyfUKIDRz9_53.339347_-6.250848@1308163237.json", "address": "78 Merrion Square", "owner": "simplegeo", "postcode": "Dublin2"}}
]}
Bring your data into PostgreSQL with something like mysql_fdw, or other methods.
Once you have this table, add a geography column, and populate the new column:
ALTER TABLE places ADD COLUMN geog geography(Point,4326);
UPDATE places SET geog = ST_MakePoint(longitude, latitude);
Now select the nearest 10 places that are within 100 kms:
SELECT places.*, ST_Distance(geog, poi)/1000 AS distance_km
FROM places,
(select ST_MakePoint(-90,47)::geography as poi) as poi
WHERE ST_DWithin(geog, poi, 100000)
ORDER BY ST_Distance(geog, poi)
LIMIT 10;
Best Answer
There appear to be 7 major "POI" databases which have a large quantity of API-accessible POI data in the US. (Cribbing much of this from http://www.quora.com/What-are-the-pros-and-cons-of-each-Places-API)
Depending on what you are looking for, these each offer different benefits.
TheWebMiner GEO
TheWebMiner GEO is a collection of POI databases. Geographical data can be downloaded on a specific radius form a point e.g. if you want data for entire United States you can select Colorado as point and 2500 miles as radius and you can get all data for United States.
Yelp, CitySearch
These two are similar sites: centered primarily around reviews, rather than acting as POI databases. As a result, their data is slanted more towards places that users would write reviews for -- you won't find "Middle East Upstairs Men's Room" in these databases. However, they tend to have more rich information -- in addition to categorization, they also have, of course, the reviews themselves, and additional information like ratings, etc. For many purposes, these review sites are superior in the places where they have content.
Facebook
The Facebook places API offers a way to integrate places which is more social in nature. Places which exist in Facebook places will tend to have things like wall posts, pictures, etc. more often, and for places which do business on Facebook, you can find more content. However, the overall quality of these places is lower; in my limited sampling, geographic accuracy tends to be lower than Foursquare, and the coverage is lower, without the benefit of curated reviews that Yelp and CitySearch offer.
FourSquare
FourSquare is the most comprehensive list of places of any of these APIs, period. It has the most comprehensive list of places -- in part because it has a more broad definition of what a 'place' is. (See above.) For venue search, it has a relatively open API, and check-ins, tips, and so on provide a reasonable set of metadata -- but more limited than the more 'refined' place sources above.
However, because much of this data is user entered, a lot of this data is crap: duplicates are a fact of life, and you will have to cope with misentered data.
Google
Google's places API provides you a limited set of information from Google's place pages. The information provided is reasonable, but it is more difficult to get information like links to places out of this API. It is a relatively comprehensive selection, but its search is not particularly broad, and it does not offer the information like checkins and so on that Foursquare offers. Overall, this API is reasonable, but with less coverage than Foursquare and (via the API) less depth than others, it doesn't offer a lot more than Foursquare or the others in either of the directions.
SimpleGeo, Factual
I haven't looked at these. However, they are 'open' data, unlike all the others, which is probably a benefit for some use cases.
Overall, which is 'best' probably depends entirely on what you're doing; they all offer different sets of points, different depth, etc. Foursquare is probably a winner on breadth; Yelp (in areas it covers) is a bigger winner on depth; Google fits inbetween them in some ways.
I'll note that this is all anecdotal; Google claims 50 million places, while Foursquare claims more in the vein of 15 million, so there may be a significant difference in the actual number of places Google has... however, at least in the mostly Urban areas I've been looking at, it seems that Foursquare is more comprehensive. (Since Foursquare data is largely user-submitted, it makes sense that urban areas will be better covered, so my observations may not hold for any given use case.)