High-integrity ETL & EDA pipeline for global ad performance. Analyzes ROAS, CPA, and CTR across TikTok, Meta, and Google Ads using Python.
Is your ad spend working, or just disappearing?" A Python-driven audit of 1,800 global campaigns across TikTok, Meta, and Google Ads. This project moves past basic reporting to find where the actual profit is hidden in the 2024 landscape.
The Mission Most ad dashboards tell you what happened. This project uses Python to figure out why. I built this to identify market inefficiencies—like the massive ROAS gap between legacy platforms and emerging channels.
The Tech Stack Python (Pandas/NumPy): For the ETL pipeline and data cleaning.
Plotly & Seaborn: For interactive, high-fidelity visualizations.
Jupyter: The full analysis playground.
Key Findings (The "TL;DR") TikTok is the Alpha: Averaging a 9.54 ROAS, it’s currently outperforming Google Ads by over 100% in efficiency.
The "Google Tax": Google drives volume but at the highest cost (CPA ≈ $48).
Winning Synergy: The "Search + TikTok" combo is the 2024 gold mine, yielding an 8.12 ROAS.
Project Structure Global Ads.ipynb: The heart of the project. Contains the ETL, EDA, and all 6 interactive visuals. Global Ads (1).ipynb
global_ads_performance_dataset.csv
The raw 2024 campaign data.
Read the Full Strategy I wrote a deep-dive on the business implications of this data over on Medium: https://medium.com/@danisinator123/the-roas-revolution-how-we-analyzed-1-800-campaigns-to-solve-the-google-tax-problem-a28bcd5abe2a
Built with precision to stop the budget leak.