jake.a.westfall@gmail.com — linkedin.com/in/jake-a-westfall

Former quantitative psychologist now working in data science/analytics.

**Statistical expertise**

- Regression-based methods — logistic regression, ridge/LASSO, splines, factor analysis, etc.
- Machine learning algorithms — clustering, dimensionality reduction, tree ensembles, etc.
- Design of complex experiments (or simple A/B tests) and power analysis
- Data visualization and dashboard creation using R Shiny
- Bayesian data analysis using Stan and PyMC3

**Technologies and programming languages**

- R, Python (expert)
- Git, Regular expressions (proficient)
- SQL, SAS, Bash scripting, Javascript, Julia (rusty but familiar)

**Open source software**

- Coauthor of Python package
**bambi**- A simple, high-level interface for fitting Bayesian multilevel models using Stan and/or PyMC3 backends
- GitHub repository stats: About 150 unique visitors, 2 unique cloners per week

- Sole author of R Shiny web app
**PANGEA**- GUI-based statistical power calculator for complex experimental designs involving any number of fixed or random factors arranged in any structure
- Site usage stats show that the app is accessed about 200 times per week

**Data analysis projects**

- What is the yearly risk of another Harvey-level flood in Houston?
- Analyzed 20 years of daily water height data from 21 stream gauges near Houston, TX
- Estimated that Hurricane Harvey caused roughly 50 – 120 year flood events at most sites, although the severity/risk varies widely depending on specific location

- Perceptions of political polarization among the U.S. electorate
- Analyzed 180,000 phone survey responses from 21,000 respondents over 40 years toward 10 political issues
- Provided evidence that most Americans tend to overestimate the actual difference in self-reported political attitudes between Democratic and Republican voters

Nov 2017 – Present: **Data Scientist**, PotentiaMetrics

May 2015 – Apr 2017: **Postdoctoral Researcher**, Psychoinformatics Lab, University of Texas at Austin

- Published peer-reviewed journal articles on statistical topics including:
- Distributed R Shiny dashboards to accompany my published papers (see Web Apps in the navigation bar)
- Served as statistical peer reviewer for 35 papers in 17 journals
- Wrote open source statistical software (see Selected Work)
- Deployed analyses and Monte Carlo simulations to Lonestar 5 supercomputer

Aug 2009 – May 2015: **Graduate Research/Teaching Assistant**, University of Colorado Boulder

- 5 years teaching lab sections of General Statistics (graduate level course, Psychology department)
- Gave 4 invited guest lectures in various graduate courses in statistics and data analysis
- Provided free statistical consulting about twice per month to researchers from diverse areas including psychology, business/marketing, linguistics, neuroscience, and ecology
- Delivered high-level talks summarizing my quantitative research at 8 prestigious academic conferences; another 13 talks to various university groups

2009 – 2015: **Ph.D., Social Psychology (quantitative focus)**, University of Colorado Boulder

- Dissertation: Optimal design of psychological experiments with multiple random factors
- Received 2015 Dosier award for outstanding psychology graduate student (award given only once per year)

2005 – 2009: **B.S., Psychology**, Oklahoma State University