Tokyo Metro Markov Models

Two things I think are pretty cool: the Tokyo subway system and Markov chains. In all seriousness, as a big proponent of learning by doing, I’m always trying to figure out how to combine some theory with real-world application, motivated by creating something cool in the process. Coming home from two weeks in Tokyo + Shanghai and a week in the Bay Area at Hack Lodge, it was a fun way to combine some things I knew about for some time (making maps, markov chains), with things I recently learned about (the Tokyo metro, the “joys” of React and front-end), and a tiny relation to a preview of stuff yet to come (Stat 212 fun…).

There’s some thinking / writing I want to get to first, but as part of this side-project thing posts for how I went about creating it will be coming up as well.

Links:

• The actual project
• Code

Related Posts:

• Lost in Japan with Reflections and a Side Project
• Making Tokyo Metro Markov Models

Technologies Used:

• Pandas - Data analysis + munging, backend transition matrix calculations
• D3 - Particle visualization and client-side movement
• React - Frontend design
• Mapbox GL - Map background