Unified Ad-Service Revenue Dashboard

The Engineering Challenge: Daily revenue reporting was fragmented across disparate ad networks (AdMob, MoPub, AppLovin), requiring hours of manual CSV merging. The business needed a unified, normalized view of earnings to identify eCPM fluctuations and optimize ad spend in real-time.

ETL Architecture & Data Warehousing:

• Automated Ingestion Pipeline:
  • Engineered a headless Python scraper suite (Selenium/BeautifulSoup) and REST API integrations to ingest daily performance metrics (eCPM, Fill-Rate, Impressions).
  • Schema Normalization: Solved the “N+1 format” problem by designing a unified MySQL schema that standardized conflicting data structures from 5+ providers, enabling apples-to-apples performance comparison.
• Scalable Data Operations:
  • Built robust, incremental ETL jobs handling ~50,000 new records/month, ensuring zero-downtime updates and historical data integrity.
• Proactive Anomaly Detection:
  • Developed a custom Python Rule Engine to monitor data stream health. It automatically triggers email alerts when critical metrics (like eCPM) drop by ≥10%, enabling immediate remediation of revenue leaks.

Business Intelligence & Impact:

• Operational Automation:
  • Replaced manual spreadsheet workflows with a self-service Tableau dashboard.
  • Impact: Reduced daily reporting overhead by >80% (cutting 2 hours of work to <15 minutes).
• Revenue Optimization:
  • The unified visibility allowed the product team to identify and remove underperforming ad units.
  • Result: Directly contributed to an ~8% increase in blended fill-rates across the portfolio.

Technologies Leveraged: Python, Selenium, MySQL, REST APIs, Tableau, Pandas, Cron, ETL Pipelines.