Chase Coleman
Data Science · Marketing Measurement · Causal Inference

Chase Coleman, Ph.D.

Team Leader, Data Science at Rocket. I work on marketing measurement, incrementality, and Bayesian modeling, and I write open-source tools for causal inference. Author of panelkit.

Get in touch GitHub LinkedIn Download CV

About

I lead the marketing measurement & incrementality team at Rocket, where I build the MMM, attribution, and experimentation systems that guide hundreds of millions of dollars in annual marketing spend across the Rocket Mortgage and Redfin brands.

I have a Ph.D. in Economics from the University of Kentucky, where I taught Ph.D.-level applied econometrics. Day to day I work on geo-experimentation and synthetic control, media mix modeling, incrementality-based attribution, and scenario planning under macroeconomic uncertainty.

Outside work I write open-source causal-inference tools. The most recent is panelkit, a Rust library for panel and geo experiments.

Experience

Jan 2026 – Present
Team Leader, Data Science
Rocket Limited Partnership · Remote
  • Lead the marketing measurement and incrementality team. Grew the testing program from about 4 toward 15–20 tests a year.
  • Made synthetic-control holdout testing the standard for measuring spend across the Rocket Mortgage and Redfin brands, using TV-viewership and clean-room data.
  • Built a Brand Health / Time-to-Impact model that links brand media to brand-health KPIs and downstream leads, which shaped upfront media commitments.
  • Launched an in-house, incrementality-based attribution system that drives CAC and profitability decisions.
  • Main data-science partner to the SVPs of Performance and Brand Marketing. Wrote the white paper behind the team's measurement playbook.
Jul 2025 – Jan 2026
Staff Data Scientist
Rocket Mortgage · Remote
  • Designed and shipped a profit-optimization algorithm that allocates spend across marketing partners, adding millions in incremental profit.
  • Led the enterprise MMM and attribution roadmap and managed a team of data scientists and engineers.
  • Built a scenario planner that pairs a Bayesian structural VAR with an LLM-generated internal codebase to produce budgets and projected P&L under macroeconomic scenarios.
  • Technical lead for causal inference and MMM across the data science organization.
Oct 2023 – Jun 2025
Senior Data Scientist
Rocket Mortgage · Remote
  • Ran geo-experiments and incrementality tests on Google and Meta to guide budget allocation.
  • Built and shipped a Bayesian Media Mix Model for annual planning and quarterly re-forecasting.
  • Built an automated value-based bidding system that cut cost per conversion by 24%.
Jul 2022 – Oct 2023
Data Scientist
Rocket Mortgage · Remote
  • Built ML segmentation models in Python and PySpark on AWS over large credit and tradeline datasets.
  • Used Bayesian methods (PyMC) to estimate milestone probabilities across the loan lifecycle, and deployed them with MLflow, Seldon, and Argo.
Summer 2021
Data Science Intern
Federal Reserve Bank of Minneapolis · Remote
  • Refactored Dodd-Frank Act Stress Test (DFAST) model code so it was easier to reproduce and maintain.
2018 – 2023
Ph.D. & M.S. in Economics
University of Kentucky
  • B.A. in Mathematics, Economics & French Literature, Transylvania University (2016).

Open Source

panelkit

Author & maintainer · PyPI · GitHub

A causal-inference library I wrote from scratch for panel and geo experiments — Synthetic Control, Augmented SC, Synthetic DiD, MC-NNM, Callaway–Sant'Anna, and Sun–Abraham — in Rust with Python bindings.

The numerical core has no dependencies (matmul, Cholesky, QR, Jacobi SVD, simplex/Frank–Wolfe solvers). It runs about 60× faster than a NumPy/SciPy synthetic control per fit, and about 1,400× faster on a full placebo test, with reproducible, bit-identical results. It also includes a geo-test design layer for power analysis, market selection, and post-test measurement.

RustPythonSynthetic ControlSynthetic DiDGeo ExperimentsPower Analysis

Research & Talks

Revisiting the Effect of Monetary Policy on Household Consumption: A Functional Approach

Ph.D. research · Python, MATLAB, time series

Estimates how monetary-policy shocks affect household consumption by identifying functional monetary-policy shocks. It uses functional local projections across the whole interest-rate term structure and finds that consumption responses vary with household balance sheets and life-cycle stage.

Monetary Policy, Interest Rate Term Structure & ConsumptionMidwestern Econometric Group; Kentucky Economics Association, 2022
Under a SALT Cap: The Effect of Limiting the SALT Deduction on Local Housing MarketsInternational Institute of Public Finance, 2022
Income and Wealth InequalityGeorgetown University, Economic Policy Academy, 2020
Invited teaching lecturesFulbright Spain & Andorra, 2017; Université Paris 1 Panthéon-Sorbonne, 2017

Teaching

Instructor for Ph.D. Applied Econometrics (University of Kentucky, 2025), a special section of Graduate Macroeconomics, R for Economists, Principles of Macroeconomics, and Economic & Business Statistics. Former Fulbright grantee (Andorra).

Toolbox

Languages & Frameworks

Python, PySpark, R, MATLAB, Rust, SQL, PyTorch, PyMC, Scikit-learn, XGBoost

Causal & Marketing Science

Bayesian MMM, geo-experimentation & incrementality, synthetic control & DiD, attribution, forecasting & time-series econometrics

Infrastructure

AWS (SageMaker, S3, Glue, Athena, Lambda), Databricks, Spark, Snowflake, MLflow, Seldon, Argo, agentic AI, data clean rooms

Languages

English (native), French & Spanish (working fluency), Japanese (N3)

Contact

The fastest way to reach me is email: chasecoleman93@gmail.com. I'm also on GitHub and LinkedIn.