Twitter provides an API for getting a place object for any given location. As per twitter API documentation at https://dev.twitter.com/docs/api/1/get/geo/search :
"Given a latitude and a longitude pair, an IP address, or a name, this request will return a list of all the valid places that can be used as the place_id when updating a status.".
API returns a place object. I am specifically interested in getting the bounding coordinates for that place which can be a Point or a polygon which is included in place object. I will then use this to check if user's location lies within some other location say New York.
Now, for any user one can fetch his profile using twitter API. API returns a json object which contains user's location which is basically a string e.g, "San Francisco, CA". This location can be passed to geo search api mentioned above to get coordinates for the polygon which defines a user's location.
I am using twitter4j library to do the same. Method 'searchPlaces()' corresponds to /geo/search API of twitter and takes a GeoQuery object as argument. However, GeoQuery object has only two constructors which take GeoLocation i.e, latitude, longitude information or ip address.
As can be seen in GET geo/search API, all parameters all optional. However, providing only two constructors in GeoQuery class forces one to provide either latitude and longitude values or ipaddress. What I have instead is the string for user's location from his profile.
Can someone please help me out with why there is this disparity between twitter4j and twitter API. Also, can someone suggest some other way to find if some twitter user is from New York.
PS: Any attempt to geocode user's location to get lat,long also requires GeoQuery object which again needs lat, long and I don't seem to be able to figure out how this can be done too.
Thanks in Advance.
I encountered this as well. A workaround is:
GeoQuery geo = new GeoQuery((String)null);
geo.setQuery("New York")