Topological Data Analysis: Information from moving Data
summary
Explains how to turn high‑dimensional point clouds into computable simplicial complexes (Čech/Vietoris‑Rips style), run persistent homology with existing packages, and output barcodes that track birth and death of topological features. Demonstrates mapper‑style pipelines on example data, shows persistent bars that survive parameter changes (interpreted as non‑noise features), and claims practical use in healthcare and engineering with working code already available.
AI-written summary · about
chapters (8)
- 0:00Topological data analysis intro
- 2:09What topology is and why it matters
- 6:20Point clouds to simplicial complexes
- 10:42Homology reveals data structure
- 16:20Persistent homology groups
- 20:12Interpreting barcodes
- 23:50Applications and real-world use cases
- 26:16TDA outperforms traditional methods
related talks
28mAnthony Bak, Principal Data Scientist at Ayasdi @ MLconf SEA
48mData Visualization in Practice
19mTopoGroups: Context-Preserving Visual Illustration of Multi-Scale Spatial Aggregates
34mVisualising Multi Dimensional Data
38mData Mapping for Data Exploration
13mGeometry for Social Good
26mAnthony Bak, Principal Data Scientist and Mathematician, Ayasdi
42mInteractive Data Analysis: Visualization and Beyond