Project detail
Sales Forecasting with PyMC
Bayesian model for estimating future sales with FastAPI and Flutter, focused on intervals and visualization.
Highlights
- Bayesian forecast of future sales
- FastAPI API + Flutter app
- Uncertainty intervals and Base64 chart
- Reusable model for commercial analysis
No public repository or demo link is published for this project yet.
Architecture
PyMC helper exposed through FastAPI and consumed from Flutter to capture historical sales and visualize the forecast
Challenges
- Tuning sampling without slowing the experience
- Explaining uncertainty in a useful way
- Keeping the API contract stable
Outcome
A practical way to estimate future sales and explore scenarios with fewer rigid assumptions
Overview
Problem
Sales projections often rely too heavily on simple averages and leave little visibility into uncertainty.
Approach
I built a Bayesian workflow to capture historical sales, infer the future distribution, and present the results in an app.
Outcome
The project enables planning with more realistic intervals and a clearer reading of the forecast.