Null Island REPRESENT!!!

[straup]

GET http://localhost:5000/?lat=0&lng=0

[ {
"midpoint_lng" : -6.699146934884048,
"name" : "Null Island",
"woe_id" : 0,
"midpoint_lat" : 1.4531262182501141
} ]

[migurski]

Wouldn’t the midpoint lat and lon be zero?

[straup]

Is it 0,0 in the shapefile?

The mid-point is being derived from the simplegeo feature class bounding box (I think...) :

https://github.com/mattb/flickrgeocoder-java/blob/master/src/main/java/com/hackdiary/geo/FlickrGeocode.java#L53

[migurski]

The shapefile might be in Mercator. I think the UN and State Department both recognize 0, 0 as the “true” midpoint of Null Island regardless of observed survey results.

[straup]

Hrm. I made sure to convert the intermediate GeoJSON file (for adding the WOE ID) from EPSG:900913 to EPSG:4326 but the center point is still:

[ {
"midpoint_lng" : -6.017946082150001E-5,
"name" : "Null Island",
"woe_id" : 0,
"midpoint_lat" : 1.3141629954000004E-5
} ]

ogr2ogr -t_srs EPSG:4326 -f GeoJSON null.json null.shp

ogr2ogr -t_srs EPSG:4326 -f "ESRI Shapefile" null_island.shp null.json

Smells like a geotools thing?

[migurski]

Those new ones are very, very small floating point values. They're legal JSON but they should just be zero.