Home

I am a PhD Economist and (currently) a Data Scientist for People Strategy and Analytics at Capital One. I love applying my ML and econometric skillset to provide stucture, insight, and analysis to fairly open-ended business problems. When I am not working, I am either spending time with my family or working on side projects (which can vary over time). I have started to use my skills to help out in environmental justice projects. I have recently compiled and submitted a demographic report of the area where the Green Ridge Landfill will be built that outlines potential environmental injustice. A copy of that most recent report can be found here. 

  
    > Joe.interests
    ["causal inference", "machine learning", "programming", 
     "data visualization", "soccer", "esports"]

    > Joe.skills
    ["python", "javascript (d3.js)", "spark",
     "R", "stata", "git", "LaTeX", "SQL"]

    > Joe.contact_me
    {
      "email": "joempatten@gmail.com",
      "linkedin": "joepatten"
    }

    > Joe.about_me
    "joepatten.com/about"

    > ▮

I have been working on visualizing data in d3.js - which I mostly use at my job (and thus cannot be shown here). I believe that there is a lot of unrealized value that comes from (creative) data visualization. The following link contains a visualization showing how rankings in the FBS (college football) have changed over the first few weeks of the 2019 season.

Projects

Below are some selected non-work projects that I have recently worked on.

Demographic Report for the Surrounding Area of the Cumberland County Landfill

This report documents the racial and age demographics of the area surrounding the proposed landfill. In the affected area, the percentage of Black residents is 77.4% higher than the Black population in Virginia. Additionally, the percent of residents under 18 is 46.6% higher than the percent of residents under 18 in Virginia.
Environmental Justice, Demographics

Community College Spending on For-Profit Enrollment

I estimate the causal impact of spending by community colleges on nearby for-profit colleges by exploiting variation of budget shocks and tuition freezes/caps by state legislatures.
Causal Inference, Two-Stage Least Squares

Uncertainty of Outcome and Superstar Effect in Esports

I evaluate the Uncertainty of Outcome Hypothesis (UOH) as well as the superstar effect on Twitch viewership in two esports: Counter-Strike: Global Offensive and Dota 2. These theories have been evalauted in many sports, but this is the first paper to my knowledge that extends the analysis to esports.
Sports Economics, Esports

Impact of Drought on Herd Composition (WIP)

Using autoregressive models, I estimate the impact of sustained drought on herd composition in Tanzania and Uganda.
Autoregressive

Esports Predictions Using Monte Carlo Simulation

Using a combination of Markov Chain and tree-based ensemble methods, I forecast the outcomes of esports matches and tournaments.
Machine Learning, Monte Carlo, Markov Chain, Esports

Active Learning to Label Data

I show the benefit of using least confident sampling and margin sampling instead of random sampling to label data in order to more efficiently label data.
Machine Learning, Active Learning

Effect of S&P 500 Inclusion on Stock Price

Using random forests, we match stocks that barely make it into the S&P 500 with those that don't. We then use difference-in-differences to estimate the effect of S&P 500 inclusion on stock price.
Machine Learning, Difference-in-Differences

Affirmative Action and Human Capital Investment

We design an experiment to mimic Affirmative Action (AA). We then estimate the effects of AA on human capital investment. Featured on FiveThirtyEight.
Experiment


Recent Activity


My Experience on the Job Market

Economics

PhD Math Bootcamp

Economics

Rocket League Esports (RLCS) Simulated Predictions

Sports Analytics

My Experience with the Google Foobar Challenge

Computer Science

Machine Learning Exploration

Machine Learning

Fun with Binary Decision Diagrams (BDDs)

Computer Science