Daniela Mendirichaga

Research Analyst

Global Development Policy Center, Boston University

May 2022 – September 2025

Three years of experimentation work with Heineken, a global market leader in alcoholic beverages. Run with the Inter-American Development Bank, the OECD and the UK Behavioural Insights Team.

• Five-arm field RCT across 3 countries. Designed and analyzed the trial in Germany, Mexico and the UK, testing menu design against low- and no-alcohol choice. The primary endpoint, millilitres of pure alcohol per basket, was modeled with negative binomial. Adoption with logistic, spend with gamma. Leading arms moved the target KPI by 8 to 20 pp.
• Six-arm advertising RCT. Tested which creative features drive substitution toward no-alcohol products: health claims, sports setting, celebrity endorsement, similarity to the parent brand. A replication then held the creative fixed across social feed and sports-bar TV, which isolates the causal effect of the delivery channel.
• Econometric simulation and estimator validation. Built Monte Carlo frameworks and synthetic datasets with known data-generating processes. Used them to validate estimator performance, benchmark causal methods against each other, and prototype an analysis before the real data existed.
• LLM-based synthetic respondents. Extended the findings to Brazil and the US, where a field experiment was infeasible, using census-calibrated synthetic respondents. Validated demographic fidelity against IBGE and US Census benchmarks, then checked the estimated treatment effects against the matching field-study subgroups.
• Pricing analytics and reproducible pipelines. Price elasticity, willingness-to-pay and demand forecasting models across markets, turned into pricing recommendations. Reproducible Python/R pipelines are what scaled the analysis across 4 trials and 10+ treatment arms.
• Advised senior Heineken leadership. Presented and defended findings across multiple engagement sessions, translating causal estimates into product, pricing and advertising strategy. Coordinated data access and paper co-authorship across all three partners.
• Global data platform, 100+ countries. Led the collection and harmonization of millions of records on alcohol consumption, pricing, regulation and scientific evidence into panel databases I designed. Built the data model, estimation framework and visualization layer behind the internal policy-intelligence platform on top of them.

Causal inference & structural econometrics

• Demand and pricing in the pizza market. BLP random-coefficients demand and supply in pyblp, instrumented for price endogeneity. Recovered marginal costs, markups and cross-price elasticities, then simulated prices, shares and welfare under other market structures.
• Sale pricing and demand dynamics replication. Replicated Shabanpour (2025) in MATLAB on a simulated 374,400-observation store-week panel of beer sales. Separated storer from non-storer demand off sale-timing variation, then reproduced the paper's demand estimates and optimal-pricing counterfactuals.
• Data centers and electricity prices. County-panel OLS testing data-center activity against US residential electricity prices.

Discrete choice and willingness to pay

• Camping park demand. Mixed logit (xlogit) recovering willingness to pay per amenity, the value of travel time, and the cross-price elasticities that show which parks compete with which.

Predictive modeling

• Airbnb booking probability, Los Angeles. Whether a listing-night gets booked or sits empty, which is the demand signal under short-term-rental pricing. Built a 1M+ listing-night panel from repeated public snapshots, derived the label from calendar availability, then benchmarked L1-logistic against XGBoost on a chronological split.
• Used-car prices. Trees and XGBoost against an OLS/Ridge/LASSO baseline, compared out-of-sample.
• Telecom churn. Logistic regression written from scratch, with bootstrap uncertainty estimates.

Forecasting

• US construction employment. Prophet forecasts with seasonality and holidays, validated by rolling time-series cross-validation.

Master of Quantitative Economics (STEM)

University of California, Los Angeles

Los Angeles, CA · September 2025 – December 2026

• AI Agents Workshop: agents, tool calling, RAG and document intelligence, multi-agent orchestration, MCP-based deployment.
• Advanced tracks in Econometrics and Machine Learning. Selective; top third of the cohort.
• Departmental Merit Award, UCLA Economics.
• Coursework: Machine Learning & Big Data for Economists; Advanced Machine Learning; Advanced Econometrics; Empirical and Mathematical Industrial Organization (the latter a PhD course); SQL & Data Management.

B.S. in Economics

Anahuac University

Mexico City · August 2017 – May 2022

• Thesis: difference-in-differences estimation of the tax-collection effects of Mexico's criminal-fiscal reform, before its repeal.
• Fellowship, Universidad Anahuac.
• Special Selection Award, CitiBanamex National Economics Prize nominee (2022).