[PolyPhy] Interactive Exploration of High-dimensional Datasets with PolyPhy and Polyglot
- Mentors
- Oskar Elek
- Organization
- UC OSPO
- Technologies
- python, javascript
- Topics
- web, data science, ui/ux, Interpretability
PolyPhy is an interpretability tool which identifies network structures in both 2 and 3-dimensional data using the novel MCPM algorithm. Users can view and rotate the resulting network. A web app, Polyglot, enables users to navigate a map of the network, hover for detailed views of points, and filter sections of the network based on a novel similarity metric. Despite its excellent interactivity, Polyglot is limited to three precomputed networks. We propose to combine the two tools such that a user can upload any of their own networks from PolyPhy and thus greatly increase the ability of users to interact with their networks from PolyPhy.