These scripts pre-process shapefile data from data.nola.gov and put it into a format suitable to be imported into OpenStreetMap.
To use, run:
make
And if everything works you should get output in the osm/
directory.
You don't need to be able to run this to participate in the import. This is available for those who are curious or would like to improve the process.
Updates to the addresses file on data.nola.gov January 2015 removed the ADDR_TYPE field from the shapefile. This causes convert.py to fail, and we depend on ADDR_TYPE to solve cases where two address points are in the exact same location.
So convert.py no longer works. For now, you can download the October 2014 shapefile from data.nola.gov by going to "more views" on the left side of this page:
https://data.nola.gov/Geographic-Reference/NOLA-Addresses/div8-5v7i
You will need at least a few GB of RAM for this to work well. We also attempt to do work in parallel, so more CPU cores will be helpful.
Python modules needed include:
- fiona
- shapely
On Linux (Debian/Ubuntu) you should make sure the following packages are installed first:
- curl
- gdal-bin
- libgdal-dev
- python-dev
- python-pip
- python-shapely
- libspatialindex-dev
Then you can build/install fiona with:
sudo pip install fiona
Based on dcbuildings and nycbuildings.
Adapted for New Orleans by Matt Toups.
See https://wiki.openstreetmap.org/wiki/New_Orleans,_Louisiana/Building_Outlines_Import
Track the progress of this import at: http://tasks.openstreetmap.us/job/41