You’ve reached Will Lowe’s homepage. I am a political methodologist specializing in statistical text analysis with applications to international relations, legislative politics, and frankly anything else that seems interesting. Formally I’m Senior Researcher at Eurodata in the MZES, part of the University of Mannheim. As of January this year I’m also part of the Collaborative Research Centre SFB 884.
Details of what I write about and the things I teach can be found in the links above or in the trusty CV. You can find also find me answering questions on stats.stackexchange.com, writing code on GitHub, R-Forge and Sourceforge, and perhaps inevitably wittering on Twitter. There’s a blog here too. Alternatively you can ponder some biographical stuff, or just a picture. Actually I look much older and more worn-out than that but you’re unlikely to get me to change it before I retire.
Our paper on ‘Estimating Central Bank Preferences’ just won the PoliInformatics NLP Unshared Task at ACL 2014.
Regrettably there have been some summer rearrangments: I won’t be teaching text analysis at the IPSA / NUS Methods summer school in Singapore.
Did you ever suspect that existing models for extracting political positions from text were basically all doing the same thing? Well, they are. Here’s some unified theory of text scaling.
If you work with event data you might reasonably wonder a) how best to wrangle it into a shape you can work with, and b) where those conflict-cooperation scales that everybody seems to use come from. I suggest the events R package for event data wrangling, and an (updated) paper on event data analysis for a way to think about inducing conflict scales and working with the resulting time series.