Project detail
Campaign Effectiveness with PyMC
Bayesian model for estimating expected conversions based on campaign spend and duration.
Highlights
- Optimization by spend and duration
- Optimal scenarios and heatmap
- Expected conversion modeling
- Useful for budget decisions
No public repository or demo link is published for this project yet.
Architecture
FastAPI backend with a Bayesian model and Flutter app to explore campaign combinations
Challenges
- Exploring grids without losing speed
- Interpreting parameter sensitivity
- Avoiding overfitting with limited data
Outcome
A foundation for allocating budget to campaigns with better expected return
Overview
Problem
Campaigns with different budgets and durations produce highly variable results.
Approach
I estimated expected conversions with PyMC to compare campaign scenarios.
Outcome
The analysis helps choose more profitable combinations before investing.