1

I'm working on a project in which I cluster locations in New York City based on their distance. My dataset contains latitude and longitude and thus I'd like to project it to a 2D Cartesian space in which distances are preserved.

I'm not really familiar with all those concepts, but I chose to use the geotools java library to project my coordinates. However I don't know to which CRS I should project in order to be able to do so, could you help me?

I've seen some cylindrical equidistant CRS but I assume the coordinates will then be in a cylindrical space. Should I then project this space on a flat surface? Is there any equidistant Cartesian CRS? How could I find a corresponding CRS in the online databases?

4
  • 2
    I would use one of the New York system planes. – Ian Turton Jun 1 '15 at 11:28
  • @iant I had a look at them but it seems that they only cover parts of New York (Long Island, North, East, West), and I need to cover the entire city and its neighborhood – Ben Jun 1 '15 at 12:43
  • 2
    Those parts are for the state - I'd guess NYC is in the East zone but Long Island isn't too far off either. – Ian Turton Jun 1 '15 at 14:03
  • 2
    Any primer on map projections will explain what a cylindrical projection is. It will also explain that equidistant projections can't possibly preserve all distances between all points. You'll need to decide if the systematic error (which over a small area like NYC, would be pretty small) is too great for your application. – Vince Jun 1 '15 at 15:17
3

New York State Plane Long Island is the state plane zone used for NYC. Its EPSG Code is 2263, units are in feet, and it uses Lambert Conformal Conic. Despite Long Island being in the name, it does cover NYC and is used by all the city agencies here that produce GIS data (Dept City Planning, DOITT, etc - for example all data provided here uses that system: http://www.nyc.gov/html/dcp/html/bytes/applbyte.shtml).

If you're measuring distances just within the city and surrounding area, that's the system to use - I believe there is an alternative one that uses meters if you prefer (EPSG 32118). Another alternative would be UTM (NYC is in zone 18N). Regardless of what you use (assuming you're not doing something that needs pinpoint accuracy), given the fact that your study area is small and you're only measuring distances within that area you're not going to get significant distortion or errors in your measurements. This only starts to become a serious issue once you're measuring greater distances across countries or continents.

A useful projection reference: http://www.radicalcartography.net/?projectionref

1
  • Thanks, I'll probably use this one (the one in meters actually), accuracy is not that important for my application as long as it is less than ~10m. – Ben Jun 2 '15 at 18:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.