Education
01Economic Quantitative
Analytics
Master of Science
DePaul University
Economic Data Analytics
Bachelor of Science
DePaul University
ECONOMIC REASONING. EMPIRICAL RIGOR.
THEOPHILE MULLER / PORTFOLIO
CHICAGO —:—Résumé
ECONOMICS · ANALYTICS · RESEARCH
Theophile
Muller.
Understanding a complex world,
through data.
I use econometrics and financial analysis to understand how shocks move markets, policy, and public opinion.
M.S. Candidate, Economic Quantitative Analytics
DePaul University
SCROLL TO EXPLORERESEARCH DRIVEN. CURIOUS BY DESIGN.
02 / CURRICULUM VITAE
A quantitative mind.
An interdisciplinary approach.
Download résumé Economic Quantitative
Experience
02Economic Graduate Assistant
DePaul University
Accounting & Data Analytics
Kehoe Design
At the intersection of economic research and real-world analysis.
THE TOOLKIT
Methods meet making.
Data & inference
R · Python · Stata
SQL · SAS
Analysis & reporting
Excel · Power BI
Build & automate
JavaScript · HTML · CSS
C# · Java
Power Automate · Power Apps
03 / RESEARCH
Questions worth asking.
Evidence worth finding.
Exploring how shocks shape attitudes,
and whether data can reveal what comes next.
APPLIED ECONOMETRICS
Terrorist Attacks and Political and Social Attitudes in Europe
A difference-in-differences study of how the 2015–2016 terrorist attacks in France affected political and social attitudes.
R / STATADownload paper
Read abstract & findings
PROBLEM
Did the attacks causally change French public opinion?
DATA
11 ESS rounds, 2002–2022 · ~322K observations · France + 13 control countries · World Bank GDP per capita, population, unemployment, and net migration · Global Terrorism Database fatalities.
METHOD
Difference-in-differences, event study, controls, and robustness checks.
TOOLS
R · Stata
RESULT
No significant immigration shift; France moved +0.256 rightward on the ideology scale.
WHAT I LEARNED
Political shocks can affect values and ideology differently, and identification assumptions matter as much as coefficient size.
FINANCIAL ECONOMETRICS
Can Machine Learning Predict Rivian Returns?
A forecasting comparison testing whether machine learning can outperform classical time-series models on daily Rivian returns.
R / MACHINE LEARNINGDownload paper
Read abstract & findings
PROBLEM
Can engineered macro and volatility features predict RIVN returns better than classical benchmarks?
DATA
613 daily observations, Jan. 2024–Jun. 2026 · 126-day out-of-sample test set.
METHOD
ARIMA · ARIMAX · GARCH family · Random Forest.
TOOLS
R
RESULT
Random Forest: best RMSE (.0305) and 64.3% directional accuracy.
WHAT I LEARNED
Nonlinear combinations of market, volatility, and momentum features carried more signal than own-return time-series structure alone.
04 / SELECTED PROJECTS
Curiosity,
made interactive.
From the world around us
to the stories that shaped it.
THE WORLD IS YOUR CLASSROOMCarte BlancheGeo.
01 / INTERACTIVE GEOGRAPHYA world of
A world of
discovery.
A browser-based geography game with multiple learning modes, daily challenges, streaks, leaderboards, and persistent player progress.
JAVASCRIPTHTML / CSSSQL
Explore Carte Blanche Geo (opens in a new tab)FRANCE HISTOIREStep inside
French history.
French history.
02 / INTERACTIVE HISTORYThe past,
The past,
in perspective.
An interactive exploration of French history with chronological navigation, historical maps, and connected events through time.
JAVASCRIPTHTML / CSS
Explore France Histoire (opens in a new tab)01 / 02TWO WAYS TO EXPLORE
05 / CONTACT
Good questions.
Great conversations.
I’m always open to discussing research, analytics, projects, and interesting opportunities.
Let’s connect