I’m at the Minnesota Developers Conference today, talking about topology and data. Hoping it’ll be fun to talk to developers about math and data! In the meantime, looking at cryptography and microservices and containerization. I’m working on trying to containerize some of my own projects…. but you might hear about that if you come to my talk.
- Here’s an older Github repository of info about TDA in R. There, launch a binder to try out the code yourself! Just press the “binder” button! This uses Docker to give you an R environment.
- Here is a link for exploring super-level sets in Dow Jones data during the period 1991-2002, leading up and through the dotcom bust. You can change the threshold (what level of correlation you want to cut off the graph edges at) and you can play with the year.
- Here is a link to an R Shiny page for exploring persistent homology in Dow Jones data during the same period, 1991-2002. Warning: it’s slow to load due to the persistent homology calculation at the beginning — give it a minute. Watch the red dots slide down and to the left as the crash happens. Why are there two groups of red dots at one point, in 2001?
If you want to learn more about using topology with big data, here are some links:
- I’ve enjoyed creating visualizations with Kepler-mapper, a Python package. In my talk, some of the images of macroeconomic data are created using Kepler-mapper. Check out the examples there — they’re beautiful.
- To make prettier mapper visualizations in R, I make heavy use of networkD3 and IGraph. IGraph also has Python bindings.
- In Python, computation of Betti numbers for analysis of the stock crash-recovery cycle was done using moguTDA.
- In R, I use TDAmapper for mapper visualization and TDA for persistent homology.
If you’re looking for other resources, my free Preparation for Math Finance course textbook is accessible and often-updated since I’m teaching right now. What else? Flash sale on coloring books 🙂