Retrieving OpenStreetMap data with Python using GeoJSON geometry as boundary [closed]

Question

I'd like to :

1. Select OSM map from a polygon geometry defined by a .geojson file

2. Extract data (ex : parks area, building height,...) included in the geometry

Any idea?

Answer 1

You can use Geopandas to read the geojson:

import geopandas as gpd

boundary_geojson = gpd.read_file('C:/yourfile.geojson')
print(bounding_geojson.head())

You can use OSMnx (https://github.com/gboeing/osmnx) to retrieve OSM data based on a polygon:

import osmnx as ox 

ox.geometries.geometries_from_polygon(boundary_geojson, tags = {‘building’: True})

If the retrieve by polygon function of OSMnx doesn't work for you (or if OSMnx wont take a geojson geodataframe as a bounding polygon) you could instead retrieve the data based on a radius around a point within your polygon. Just give the coordinates of a point roughly in the middle of your geojson polygon and a sensible radius in metres around the point that you know will include all of the necessary data you will need (below example is for buildings data):

import osmnx as ox 
import ast

point = 'point coordinates e.g 51.93802486000275, 4.361977332600682' 
dist = 'distance in m e.g 2000'
buildings = ox.geometries.geometries_from_point(point, {'building': True}, dist=dist)

And convert to a geodataframe:

buildings_save = buildings.applymap(lambda x: str(x) if isinstance(x, list) else x)

Then clip the buildings data to your boundary geojson file using geopandas:

buildings_final = gpd.clip(buildings_save, boundary_geojson)

Plot the data to check if you want:

import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize = (15,12))
buildings_final.plot(ax = ax, color = 'red', edgecolor = 'black',)
plt.title('Buildings Data')
plt.show()

Save the files to geojson or shapefile if you want:

buildings_final.to_file("C:/output.geojson", driver="GeoJSON")
buildings_final.to_file("C:/output.shp")

This retrieve data by point method is a bit clunky but works. For more info about retrieving OSM data based on geometries check out https://osmnx.readthedocs.io/en/stable/osmnx.html#module-osmnx.geometries