PyData Global 2023

Hands-On Network Science
12-08, 20:00–20:30 (UTC), General Track

In this talk, we will introduce network science and demonstrate its usefulness in mining different types of data, including social network data, time series data, and spatiotemporal data. Our talk will include practical, hands-on examples of real-world problems we've solved in the developing world with tools from network science--including epidemic forecasting, stock market crash prediction, and food pricing trend analysis across regions. Python code will be available for those who want to run the analysis themselves.


This talk is for intermediate and advanced data scientists hoping to add new tools to their analytics. We'll quickly overview the basics of network science (3 minutes) before diving into three case studies that apply tools from network science to real-world data science problems (7 minutes each).

Each case study will present the problem (epidemic spread, stock market prediction, and food pricing trends), overview the data, and detail the methodology we used to solve the problem. Code and data will be available open-source for participants interested in running the analysis or similar analyses themselves.

By the end of the talk, participants will be able to analyze data at scale using open-source network science tools rather than computationally-costly traditional methods for time series and spatial data problems.

Ideally, participants will have some background in Python programming and knowledge of basic time series and spatial data analytics. A background in network science is not necessary.


Prior Knowledge Expected

Previous knowledge expected

Colleen Farrelly is a principal data scientist whose expertise spans network science, topological data analysis, quantum computing, and natural language processing. She and Dr. Gaba have a first book, The Shape of Data, which covers many network science tools.

Dr. Yae Gaba is a researcher at Quantum Leap Africa whose expertise includes computational geometry, topology, graph learning algorithms.

Dr. Franck Kalala Mutumbo is a researcher at African Institute of Mathematical Sciences and University of Lubumbashi. He is an expert in network science and has trained many African network science researchers. He and Ms. Farrelly have a network science book coming out in 2024.

Franck Kalala Mutombo is a Professor of Mathematics at Lubumbashi University and former Academic Director of AIMS-Senegal. He previously worked in a research position at Strathclyde University and at AIMS-South Africa in a joint appointment with the University of Cape Town. He holds a Ph.D. in Mathematical Sciences from the University of Strathclyde, Glasgow, Scotland. He is an expert in the study and analysis of complex networks structure and applications. The most recent study considers the impact of network structure on long-range interactions applied to epidemics, diffusion, and object clustering. His research interest includes Differential Geometry of Manifolds, Finite Element Methods for PDEs, Networks, and Data Science.

As a mathematician, I have a keen interest in both theoretical concepts and practical applications, particularly in modeling real-world problems that are relevant to education and industry.