Multiple Source Isochrone Map Generator

Create your own isochrone maps fast with the buffered roads method from OpenStreetMap road networks.

Results

Obtain your own graph

There are two ways you can procure a graph.

First way

Run the Python code in osmnx_getmap.py. Change source_coord to the coordinate of interest, dist to change the network size, network type to 'bike' or 'driving'.

import osmnx as ox
ox.config(log_console=True)

source_coord = (20.148722, 46.252382)
G = ox.graph_from_point(source_coord, dist=100, network_type='bike')
ox.save_graphml(G, 'out.graphml')

This uses the internet to download the graph. File out.graphml is ready to be used.

Second way

Download a map from the website of GeoFabrik in .osm.pbf extension (e.g. someplace.osm.pbf). Copy the coordinates of a polygon in a text file (e.g. poly.txt) in the way presented in osmosis wiki and save. Run the following command.

osmosis -verbose --read-pbf file=someplace.osm.pbf --bounding-polygon file=poly.txt
completeWays=yes --tf accept-ways highway=* --tf reject-ways highway=footway
--tf reject-relations --used-node --write-xml file=extract.osm

See more details of osmosis here. If the polygon stretches beyond multiple territories consider using the merge over multiple .osm.pbf files. Run the Python code in osm2graph.py where extract.osm is the file created by osmosis:

import osmnx as ox
G = ox.graph_from_xml('extract.osm', bidirectional=True)
ox.save_graphml(G, 'out.graphml')

File out.graphml is ready to be used.

Compute the isochrones

The program is designed to be faster than networkx/ shapely in Python. It makes use of various boost libraries for fast computation: Boost Graph Library for graph algorithms, Boost Polygon Library for unary union of polygons, Boost Geometry for buffering polygons, Boost Program Options for parsing command line arguments. Compile isochrones.cpp with

g++ -std=c++11 isochrones.cpp -o isoch -O3 -L /usr/lib/x86_64-linux-gnu/ -lboost_graph -lboost_program_options

note that the location of the lib files might differ from /usr/lib/x86_64-linux-gnu/

When running the program make sure you have the source points in a WKT Linestring in a text file.

Isochrones with filled interior

Run with --fill on e.g.

./isoch --simplify 0 --travel-speed 20 --radii 5 10 15 20 25 30 35 40 45 50 55 60 --circle-points 11 --verbose --subtract --buffer-distance 0.005 --fill --input-file in.graphml

Isochrones without filled interior

Run with --fill off e.g.

./isoch --simplify 0 --travel-speed 20 --radii 5 10 15 20 25 30 35 40 45 50 55 60 --circle-points 11 --verbose --subtract --buffer-distance 0.00006 --input-file out.graphml

Isochrones of the shortest distance tree

Run with --predecessor on e.g.

./isoch --simplify 0 --travel-speed 20 --radii 1 2 3 4 5 6 --circle-points 11 --verbose --subtract --buffer-distance 0.00002 --input-file out.graphml --predecessor

Visualisation

We use folium in Python to create the maps. Just run python3 iso_folium.py.