Case Studies

Proof that I can turn data into decision support.

These examples show the consulting value behind my work: cleaner reporting foundations, clearer operational metrics, better forecasting workflows, and business-facing analytics outputs.

Enterprise Operations Intelligence Platform

I built this as an end-to-end analytics engineering and BI workflow for the kind of problem many growing teams face: raw files exist, but the data is not ready for trusted reporting. The project moves from CSV cleaning and validation into PostgreSQL, dbt transformations, dimensional modeling, data quality checks, and dashboard-oriented SQL for operational and product analysis.

Decision: Give operators a trustworthy view of product and operational performance from raw files.

Data: CSV inputs cleaned and validated before loading into PostgreSQL and dbt models.

What I built: Dimensional models, data quality checks, and dashboard-ready SQL for analysis.

Result: A repeatable BI foundation that shows how messy source files can become decision-ready reporting.

What to watch next: Refresh reliability, metric ownership, and whether stakeholders act from the same definitions.

Python PostgreSQL dbt SQL Tableau

Havi Hackathon - Demand Forecasting

This project was built for the HAVI / NIU Quick Service Restaurant Demand Forecasting Challenge, where the business problem was predicting daily menu-item demand by restaurant. The workflow combined exploratory analysis, time-based validation, calendar, weather, promotion, and historical demand features, plus gradient boosting models designed around the competition's wMAPE objective.

Decision: Improve demand planning for daily restaurant menu-item volume.

Data: Historical demand joined with calendar, weather, promotion, and restaurant-level context.

What I built: A time-aware validation workflow and gradient boosting models aligned to wMAPE.

Result: A forecasting approach that connected model evaluation to a supply chain planning question.

What to watch next: Forecast drift, promotion effects, item-level bias, and planning adoption.

Python LightGBM scikit-learn Forecasting Supply Chain Analytics

Ecommerce Strategy Analysis - A/B Testing

This project analyzes an ecommerce A/B test by cleaning experiment data, validating the control and treatment setup, and comparing conversion behavior and session engagement across groups and countries. It demonstrates how experimentation work should connect data validation, SQL analysis, Python workflows, and business interpretation before a team acts on test results.

Decision: Understand whether an ecommerce test produced a credible conversion signal.

Data: Experiment assignments, sessions, country segments, and conversion behavior.

What I built: A Python and SQL workflow for cleaning, validation, group comparison, and interpretation.

Result: A clearer readout of test behavior before turning analysis into a business recommendation.

What to watch next: Sample quality, segment-level effects, test duration, and repeated learning across future tests.

Python SQL Pandas A/B Testing Statistical Analysis
More Work

Additional examples across analytics and consulting

More examples that show the range of problems I can support: engagement analysis, business strategy, predictive modeling, dashboarding, and application workflows.

Group Projects

Solo Projects

FAQ

How to read these case studies

What's the difference between an analytics audit and a dashboard build?

An analytics audit identifies the decision, metric gaps, data quality issues, and reporting workflow problems before a team invests in a build. A dashboard build creates the reporting asset itself, including cleaned data, visual views, KPI definitions, QA checks, and handoff notes.

Can you analyze Adobe Target / A/B test results?

I can analyze A/B and multivariate tests, including Adobe Target and AB Tasty contexts, by checking setup quality, traffic splits, audience logic, KPI movement, segments, and business implications.

Last updated July 6, 2026.