Industry Projects

📊 Industry Analytics
Marketing Optimization Project
Informed client strategy for budget plan by evaluating the market budget plan on customer behavior across 500 locations. Locations could opt-in to the budget plan, so randomization was vioalted. I needed to estimate the causal impact of the budget. Determined the budget plan did not positively affect KPIs by performing propensity score matching and regression analyses in R, and visualizing findings in ggplot2 to deliver actionable insights to client.
R Propensity Score Matching Regression Quasi-Experimental Design Client Facing
The Question Did the new budget plan positively impact KPIs (overall sale and giftcard sales) during Q4?
What I Did Defined KPIs and detrimined analysis plan. I inherited "matched controls" that used propensity score matching to match the locations that opted-in to the budget plan. I ran multi-variable regression to see how the budget plan impacted KPIs after controlling for baseline factors (which I decdied which baseline factos to include).
Impact & Metrics Detrimined the budget plan did not positively increase KPIs after controlling for baseline factors. While there was no overall impact of the budget, there were some interesting regional differences in how the budget was affecting KPIs. I presented these findings to the client and recommended that if they wanted to proceed with the budget roll out, they should focus it on regions where it looked like it made an impact on sales.
📊 Industry Analytics
Product Benchmarking Analytics Project
Informed product strategy and technical roadmap by leading a team of 2 data analysts in collaboration with the Product team to benchmark an internal Python package against industry-standard software, and presenting findings to Research and Executive Leadership.
Python Technical Benchmarking Cross-Functional Collaboration
The Question How does our custom Python package compare to industry-standard software for analzying actigraphy (sleep) data? How should we pitch this package to expert researchers in the field?
What I Did Led a small team in to determine the analysis plan and how to best benchmark our product. I devised our analysis plan, visualization, and final presentation. We demostrated that our product was superior to industry-standard software by showing that our product could identify more nuanced sleep windows.
Impact & Metrics Presented final presentation to Executive and Research Leadership to showcase how specifically our product was superior to the industry-standard software. The insights from this presentation where used as the basis for how to pitch the product to a broader field of researchers.