This is an update of a benchmark of popular graph / network packages post. This study aims to serve as a starting point for anyone interested in applied graph or network analysis. The featured network packages offer a convenient and standardised API for modelling data as graphs and extracting network related insights. Some common use cases include finding the shortest path between entities or calculating a measure of centrality such as the page rank score. [Read More]
Efficient Large Graph Propagation Algorithm
Cross-posting from my company’s blog, but if you have not checked it out, I wrote an interesting technical piece on how we engineered a large scale label propagation algorithm.
The accompanying slides can be found here
Benchmark of popular graph/network packages
This post is superseded by an updated benchmark In this post I benchmark the performance of 5 popular graph/network packages. This was inspired by two questions I had: Recently, I have been working with large networks (millions of vertices and edges) and often wonder what is the best currently available package/tool that would scale well and handle large scale network analysis tasks. Having tried out a few (networkx in Python and igraph in R) but on different problems, I thought it would be nice to have a head to head comparison. [Read More]
Cleaning openstreetmap intersections in python
Introduction It has been a while since I have posted anything on Python, so I thought it is time to switch things up and write do a Python GIS tutorial. GIS in python typically revolves around the geopandas and shapely packages. If you are using OpenStreetMaps(osm) in your work, the osmnx package is also very useful and makes downloading and visualising map data straightforward. In this post, I explore the problem of simplifying route intersections. [Read More]