Social Network Analysis with Python and NetworkX

Materials for the PyData Barcelona 2017 workshop on Network Analysis with Python and NetworkX.

Abstract

Social Network Analysis (SNA) has a wide applicability in many scientific fields and industries. This workshop is a gentle introduction to SNA using Python and NetworkX, a powerful and mature python library for the study of the structure, dynamics, and functions of complex networks. Participants in this workshop should have a basic understanding of Python, no previous knowledge of SNA is assumed.

For this workshop attendees will need to install NetworkX (>=1.11), Matplotlib (>=1.5), numpy (>=1.10) and have a working Jupyter Notebook environment. Some examples will also use Pandas (>=0.17) and Seaborn (>=0.7), but these packages are not essential. Only basic Python knowledge is assumed.

Outline of the workshop

1. Brief Introduction to Graph Theory

   • Mathematical foundation of Social Network Analysis.
   • Why graphical representations usually doesn't help much.

2. Creating and Manipulating Graphs

   • Data Structures: Graphs, DiGraphs, MultiGraphs and MultiDiGraphs.
   • Adding nodes and edges.
   • Adding and updating node and edge attributes.
   • Graph generators.
   • Visualizing graphs using Matplotlib.
   • Common formats for reading and writing Graphs.

3. Network Analysis

   • Basic concepts: Degree.
   • Distance measures: paths, simple paths, and shortest paths.
   • Node centrality analysis: measures and their relation.
   • Analyzing groups and subgroups: Cliques, k-cores, components, and k-components.

4. Bipartite Graphs

   • Definition of bipartite networks and their use in modeling group affiliations.
   • Working with bipartite networks with NetworkX.