Avatar

Philipp Hunziker

Quantitative Researcher

Google

Biography

I’m a Quantitative UX Researcher at Google (and YouTube), where I develop experiment infrastructure, analyze logs data, and do some statistical programming.

Prior to my current role, I was a Postdoc at Northeastern University’s Network Science Institute/Lazerlab. I was also a Postdoctoral Affiliate at Harvard University’s IQSS. I hold a PhD from ETH Zurich, where I was part of the ICR group, as well as the ETH Risk Center.

Interests

  • R Programming
  • Bayesian Optimization
  • Geospatial Machine Learning

Education

  • PhD in Political Science, 2015

    ETH Zurich

Recent Posts

Modeling Swiss Toponyms with NLP

In this post I teach a neural network the linguistic and geographical patterns of Swiss place names (‘toponyms’). The resulting model is able to generate new place names for a given location, predict the most likely location of an arbitrary (and potentially fictional) place name, and allows us to explore the geographic patterns underlying the distribution of Swiss place names.

Run a Headless Mopidy Spotify Server on Ubuntu 18.04

I just spent way too long trying to set up a Mopidy audio server on my personal headless Linux machine, so I thought I might as well write down what I learned for posterity.

Projects

*

arscpp

arscpp is a black-box Adaptive Rejection Sampler implemented in R and C++.

velox

velox is an R package for performing fast extraction and manipulation operations on geospatial raster data.

BoostLines

BoostLines is an R package providing highly scalable geometric line operations for planar networks.

MapColoring

MapColoring is an R package implementing graph-theoretic optimal contrast coloring of polygon maps.

pgrid

Convenient access to PRIO GRID raster data from R.

Recent Publications

Geospatial Conditional Random Fields: A Probabilistic Model for Explaining Spatial Tesselations

This paper introduces a probabilistic method for modeling (geo-)spatial tesselations. We propose a new architecture of conditional …

Scalable Spatio-Temporal Autoregressive Models for Large Non-Gaussian Multivariate Data

Very large spatio-temporal lattice data are becoming increasingly common across a variety of disciplines. However, estimating …

Accounting for Immunity in Spatial Processes: The Spatial Susceptibility Model

Political scientists frequently study spatially interdependent processes, such as policy diffusion, democratization, or the spread of …

Roads to Rule, Roads to Rebel: Relational State Capacity and Conflict in Africa

State capacity is often described as one of the most important explanations of civil conflict. Yet current conceptualizations of state …

New Spatial Data on Ethnicity: Introducing SIDE

Research on ethnic politics and political violence has benefited substantially from the growing availability of cross-national, …