The title of this post is a little bit of bullshit. This summer, I intend to read through the book Random Geometric Graphs by Mathew Penrose, sort of fumbling my way through it and discussing along the way with my advisor. On the garden path, it occurred to me that one might be able to use point processes or random fields to model sound change in language. One could possibly apply these concepts to events like the Northern Cities Vowel Shift to inform were our sounds will go, based on where they have come from. It’s a pretty half-baked idea right now, but the parallels seem present. A starting point might be this here book.
Regarding the redistricting portion of the title, I recently read an article on Wired about redistricting and in the process discovered the work of Wendy K. Tam Cho, who uses simulation strategies to measure the partisan advantage induced by a particular redistricting map. Such a simulation procedure is subject with a multitudinous number of constraints, due to the necessity of contiguous districts and trivial things like the Voting Rights Act. But clearly, this is another usage of a spatial stochastic model to solve novel and thus far intractable real-world problems. To both these ends, I have more ideas than intent or even ambition to follow through on them.
To whom it may concern,
I have been gradschooling for the last 6 months or so, so I have had nary the chance to freelance and kick shit around as in the before-times. However, over break while studying for qualifying exams, I put together a few solutions for “An Introduction to Generalized Linear Models” by Dobson & Barnett. You will find it here (let me know if there are errors).
Clearly, I’d like to get more stuff out here but I’m NOT a content-producing machine. This semester I’ll be taking some classes on stochastic processes, linear models and the like so maybe I’ll have something of that ilk pop up here. Probably not, but like the Tunguska event of 1908, in which a large explosion was reported of what purportedly was a giant, 200 foot meteor, which had energy 1000 times greater than when Little Boy was dropped on Hiroshima, yet mysteriously left no impact crater, strange things can happen. Ok.
Well, it’s that time. I’m going to grad school soon to start my PhD in Statistics at the lovely Purdue University, along the banks of the Wabash. Clearly I’ve done stuff in the past few months and my lack of blog posts is a bad indicator of my interstitial life experiences. I finished my job at UnitedHealthcare, went to Paris and Amsterdam, worked grading standardized tests of the stellar Louisiana education system and started an internship at MatrixCare — a company that provides software to make life easier for long-term care providers. My experience at MatrixCare has been great and I’ve learned a ton about data science in a real world environment. It’s been great.
To sign off, I’ll give a visualization that I produced while I was practicing the
vcd library in R and messing around with the fascinating GSS Data Explorer.
It’s a faceted mosaic plot showing the association between bilingualism and life outlook, with separate panels for the biennia between 2006-2014. It’s fairly interesting to see how bilinguals are more highly allocated into exciting outlooks and less allocated in the routine/dull category with the trend reversed for non-bilinguals. For one, my life would be less interesting had I not heard of De Jeugd Van Tegenwoordig or Downistie, a couple of Dutch classics. Speaking of which, follow me on if you’re on there. I’ll get the code up eventually.