Tsangares / point_to_polygon_distance

Quickly get the distances between a set of points and a set of polygons. If the polygons are concave, the program will simply use the convex hull of the polygon to approximate distance.

Summary

This code will take coordinates and polygons and produce a matrix of distances from each coordinate to the convex hull of the polygon. This method is quick but loses some precision on extremly non-convex polygons in the input shape file.

Quickstart

Navigate to the working directory and make a virtual environment.

python virtualenv env

Activate the virtual env

. env/bin/activate

I use fish so,

. env/bin/activate.fish

Then install the pip dependencies,

pip install -r requirements.txt

Navigate to ./src

cd src

Then run get_distances.py

python `get_distances.py

After about 4 min for 600,000 samples with 7 processes, it will produce a parquet in ./data/dist/ that is a matrix of distances from point to convex hull of the polygons in the given shape file. To match it back to original samples, load it into memory then match its rows to points and columns to polygons. The associated points and polygon files will be csv files also in ./data/dist.

To test on a sample change sample to true on the line:

traffic_stops = get_traffic_stops(sample=False)

To make:

traffic_stops = get_traffic_stops(sample=True)

And the program will run on a sample of 10,000.