Topological-Data-Analysis on Ryan Orban

Topological Data Analysis from Ayasdi

me@ryanorban.com (Ryan Orban) — Wed, 04 Dec 2013 00:00:00 +0000

Summary

Ayasdi was a Stanford spinout (founded by Gunnar Carlsson, Gurjeet Singh, and Harlan Sexton) that commercialized topological data analysis (TDA) — specifically the Mapper algorithm, which Carlsson and Singh had developed to extract shape and structure from high-dimensional datasets without requiring a fixed number of clusters or a distance metric.

The Mathematical Shape of Big Science Data

me@ryanorban.com (Ryan Orban) — Sat, 05 Oct 2013 00:00:00 +0000

Summary

Quanta Magazine’s coverage of topological data analysis (TDA) explained how mathematicians were applying algebraic topology — specifically persistent homology — to find structure in high-dimensional datasets where conventional clustering and dimensionality reduction techniques fail. The company Ayasdi, founded by Stanford mathematician Gunnar Carlsson, was commercializing these methods.

Ayasdi: Automatic Insight Discovery

me@ryanorban.com (Ryan Orban) — Mon, 21 Jan 2013 00:00:00 +0000

Summary

Ayasdi was a Stanford spinout commercializing topological data analysis (TDA) for enterprise use. The company was co-founded by Gunnar Carlsson (Stanford mathematics professor and co-developer of the Mapper algorithm), Gurjeet Singh, and Harlan Sexton. The product was Iris — a platform for finding hidden structure in high-dimensional datasets without requiring analysts to specify what to look for in advance.

Data-Visualization Firm's New Software Autonomously Finds Abstract Connections

me@ryanorban.com (Ryan Orban) — Mon, 21 Jan 2013 00:00:00 +0000

Summary

Ayasdi launched Iris, a topological data analysis (TDA) platform, in early 2013. The pitch was unusual: instead of building dashboards to visualize data that analysts already understood, Iris used algebraic topology to find structure in high-dimensional datasets automatically — connections and clusters that analysts wouldn’t have known to look for. The Wired piece framed this as “autonomously finding abstract connections,” which was both accurate and somewhat misleading.