Md Samshad Rahman

Md Samshad Rahman

He/His

MSc Applied CS grad Dalhousie University 🎓 | Python backend | LLM fine-tuning | AWS/GCP/Docker | Data-driven problem-solver

Unified Ad-Service Revenue Dashboard

During my time at Free Pixel Games Ltd., I identified a critical need: a clear, consolidated view of daily earnings across all our ad networks. My solution was to build a comprehensive tool that answered the simple but vital question, “How much did we actually earn today, across every single ad network?” – and to do it automatically.

Key Contributions & The Impact I Made:

  • Automated Multi-Network Data Ingestion:
    • I engineered robust headless Python scrapers (using Selenium and BeautifulSoup) and utilized REST APIs to automatically pull daily earnings, eCPM, fill-rate, and error data from numerous ad networks including AdMob, MoPub, AppLovin, AdColony, and others.
    • A key challenge I overcame was standardizing the various disparate CSV formats from these networks into a single, common schema for consistent analysis.
  • Centralized & Efficient Data Warehouse:
    • I designed and implemented a normalized data structure in MySQL. This central warehouse enables fast data joins, allowing for quick cross-network comparisons and trend analysis.
    • I built incremental ETL jobs that reliably append approximately 50,000 new rows of data each month with zero downtime, ensuring the dashboard always has the latest information.
  • Single-Page Analytics Dashboard for Clear Insights:
    • I created a comprehensive, single-page dashboard using Tableau (with an Excel Power Query version as a fallback) to display revenue, impressions, fill rate, and network health metrics side-by-side.
    • The dashboard features conditional formatting to instantly flag under-performing ad placements, enabling quick action.
  • Proactive Alerting & Trend Identification:
    • I developed a simple yet effective Python-based rule engine that automatically emails alerts for anomalies, such as a significant (e.g., ≥10%) drop in eCPM.
    • The visualized trend lines in the dashboard clearly revealed seasonality in ad performance, directly informing strategies for ad unit optimization.
  • Significant Operational Improvements:
    • This automated solution drastically reduced manual spreadsheet work by over 80% (slashing daily effort from two hours to less than 15 minutes).
    • Empowered by these clear insights, the product team was able to optimize ad spend and successfully increased the blended fill-rate by approximately 8%.

What This Project Showcases About My Skills:

  • End-to-End Data Solution Development: From automated data extraction and warehousing to insightful visualization and alerting.
  • Automation & Efficiency: Proven ability to identify inefficiencies and build solutions that save significant time and resources.
  • Data Engineering & Warehousing: Expertise in web scraping, API integration, data cleaning, schema design (MySQL), and ETL processes.
  • Business Intelligence & Visualization: Proficiency in creating actionable dashboards with tools like Tableau and Excel Power Query that drive decision-making.
  • Problem-Solving & Initiative: Taking ownership of a business challenge and delivering a high-impact solution.

Key Technologies I Mastered & Applied: Python, Selenium, BeautifulSoup, REST APIs, Pandas, MySQL, Tableau, Microsoft Excel (Power Query), Cron, ETL processes.