Network Analytics and Visualization in Python

Welcome to Section 2!

Networks encode topological information. In order to place them across the plane, for instance, on a 2D screen, we need to apply so-called layout algorithms, which, based on various logic, compute the exact set of coordinates each node should be placed on. In this video, we will overview how to compute and use such layouts in Python using NetworkX.

Network centralities are one of the most essential metrics when it comes to network analytics - they quantify the importance of network nodes from various perspectives. In this video, we are going to learn how to compute such centralities and measure node importance.

Summarizing a few of the previous NetworkX and Python commands and adding a few new ones, in this video we learn how to conduct a statistical characterization of an input graph.

In this lecture, we are going to briefly overview and practice what network communities are and use NetworkX to extract such communities from a test network. Then, we visualize the network, highlighting the different communities.

In this brief lecture, we are going to overview all the possible ways we learn to customize graph visuals using networks and matplotlib, including various node and edge characteristics.

Welcome top Section 3!

In this section, we will explore how to create synthetic networks using NetworksX. The ability to create such networks, following famous network models, will help to generate various reference networks when analzing real networks.

Note: In NetworkX 3.6+, replace random_tree() with random_labeled_tree().

In this lecture we are going to re-visit the Witcher network already covered in my Network Visualization introductory course. Here, we start from the raw .gexf data file storing this social graph, connecting every significant Witcher character based on their shared storylines and interactions. In this lecture, using this graph, we will conduct a series of analytical and exploration steps and learn how to do more elaborate visualizations in networkx, putting the learning outcomes of the previous chapter to a more advanced level. While doing so, we will also cover further, more advanced concepts of network analytics.

When reporting various measured values, such as network diameter, of a real graph, the question often arises - compared to how we should interpret those values. This question is the easiest to answer by comparing our network to randomized null-model graphs. Hence, in this lecture, we will analyze three real-life networks (all introduced in the previous course: The Witcher Social Graph, a network connecting Python packages, and the network of Wikipedia-famous personnel of the Manhattan Project). Namely, we will analyze these networks and their randomized counterparts in a small data pipeline.

This video is for those of you who are already using Gephi to visualize networks. In this lecture, we are going to revisit how to export node and edge tables from Gephi. Then, we will learn how to read and use them in Python, including how to create new node and edge attribute columns. Then, we learn how to channel these back into Gephi to use for further graph visual customizations.

Congratulations on finishing this course, and thank you for being here!