Social Network Analysis in Python
summary
Demonstrates using Python (mostly NetworkX) with dict-of-dicts adjacency, sparse matrices and edge/node objects, plus export to D3 for browser visualization and experiments with Neo4j. Shows Dijkstra/Fibonacci-style shortest paths but speeds median path-length estimation with approximate median sampling (sample sizes on the order of 1,000 even for much larger graphs). Compares attack strategies: removing high betweenness or degree quickly fragments networks while random removals do not, and cautions that claimed power-law degree fits often need better statistical testing.
AI-written summary · about
chapters (11)
- 0:00Social network analysis intro
- 1:03What is a social network
- 9:27Graph representation in Python
- 19:58Node properties in networks
- 29:40Average shortest path length
- 32:34Clustering coefficient and transitivity
- 37:21Degree distribution and power laws
- 48:18Generative models for networks
- 54:03Centrality metrics and robustness
- 1:00:12Visualizing networks with JavaScript
- 1:05:18Q&A on graph formats
related talks
26mGetting an Edge with Network Analysis with Python
23mD&D and G a daring tale of Dungeons and Dragons and also Graph
37mGraphs, Networks and Python: The Power of Interconnection
01h 06mComplex networks sparsification: methods and tools
28mNetworkX is Fast Now: Graph Analytics Unleashed
38mThe NetworkL python package
01h 25mNetwork Science, Game of Thrones and US Airports
28mImproving Efficiency in Transportation Networks using Python