Large, high dimensional data sets are generated in several areas of research including medicine, ecology and social science, and can be difficult to analyse and visualise due to their scale and complexity. Topological data analysis is a useful approach that can reduce the dimensionality of a data set while preserving its key features.
The Joining the Dots project aims to develop an analytical workbench based on these techniques, generating network maps of the links between health metrics and outcomes in large groups.
The RSG was approached by researchers in Maths and Bioinformatics to develop a web interface for a data management, analysis and visualisation framework based on Kepler Mapper, an existing Python library for machine learning network analysis.
In a follow-on project, we are also working to add further features for the topological analysis of data sets, including persistence homology analysis using the GUDHI Python library.