Paul Rosen
Abstract: Common information visualizations, e.g., scatterplots, suffer from overdraw even with modest amounts of data. Several techniques exist to reduce this overdraw, e.g., changing visual encodings or subsampling data. However, most guidance on their use remains largely rules-of-thumb. By applying Topological Data Analysis (TDA) to the problem, we have developed techniques that are mathematically robust, correspond to human perception and cognition, and are surprisingly effective at selecting effective visualizations of data. This tutorial will introduce participants to techniques for resolving these issues on three common information visualizations, namely scatterplots, line charts, and graph visualizations. The solutions to these problems are a mix of optimization interfaces and mechanisms for interactively exploring data, all using Topological Data Analysis.
Topological Data Visualizaton Workshop
May 16 - 20, 2022
University of Iowa
https://homepage.divms.uiowa.edu/~idarcy/CONF/TDV.html
This file contains 3 demos and associated datasets:
- TopoClusters: Topologically-based Scatterplot Optimization
- TopoLines: Topologically-based Line Chart Smoothing
- UntangleFDL: Topologically-based Methods for Graph Layouts and Interaction
Requirements:
- Python3 (https://www.python.org/downloads/)
- Python Virtual Environments (should be included in Python3; https://docs.python.org/3/tutorial/venv.html)