Project Overview
During my professional tenure at Publicis Health, I identified a massive inefficiency in how weekly reporting decks were generated for executive leadership. I took the initiative to engineer a completely autonomous software solution to handle the repetitive data wrangling, transforming a multi-day manual process into a seamless, automated workflow.
The pipeline was constructed using a diverse tech stack. It began with deeply optimized SQL queries that extracted real-time performance metrics from enterprise data warehouses. This raw data was then ingested by custom Python scripts utilizing the Pandas library to execute complex data cleaning, normalization, and statistical grouping instantly.
The final mile involved transferring this highly structured data into executive-ready visualizations. I developed macro-enabled VBA templates that bridged the gap between Excel analytics and final PowerPoint presentations. This end-to-end automation dramatically reduced human error and cut deck generation time to just 5 minutes.
Key Contributions
- Engineered robust Python scripts utilizing Pandas to clean, format, and aggregate complex datasets automatically.
- Wrote optimized SQL queries to extract data efficiently from enterprise databases directly to local staging environments.
- Developed macro-enabled VBA scripts connecting Excel analytics with PowerPoint, dramatically reducing manual deck generation time to just 5 minutes.