Project
Multi-Channel Marketing Analytics
Unified KPI model across Facebook, Google Ads, and TikTok.
PythonPandasStreamlitPlotly
3 (FB / Google / TikTok)
Channels
CTR · CPC · CPA · ROAS · CPM
KPIs modeled
Streamlit Cloud
Deploy target
The problem
Marketing teams running campaigns across Facebook, Google, and TikTok get three different schemas, three different KPI definitions, and three different timezone conventions. Comparing cross-channel performance in that state is guesswork.
What I built
An analytics-engineering layer that standardizes the three platforms into one model and an interactive dashboard on top:
- Unified model: Column-level standardization (spend, impressions, clicks, conversions, revenue) across all three channels.
- KPI library: CTR, CPC, CPA, ROAS, CPM computed consistently — with clear formulas in code.
- Dashboard: Streamlit + Plotly — real-time filtering, comparative channel analysis, trend views.
- Data quality: Dedicated QC script validates dashboard numbers against source CSVs; mismatch → fail loud.
- Deliverables: CSV export + auto-generated campaign recommendations.
Why it matters
Shows the analytics-engineer half of my title: business-facing metrics with defensible definitions, not just pipelines moving bytes. The QC script is the tell — real analytics engineering is as much about trust as it is about SQL.
Questions about this project?
Email me