Looking for feedback on architecture: forecasting + optimization pipeline

[chripiermarini]

Hi all,

I built an end-to-end decision intelligence pipeline combining:

• demand forecasting (baseline models)
• model selection (WAPE-based)
• optimization (OR-Tools transportation problem)

The system aggregates forecast into average daily demand before optimization.

I’d love feedback on:

1. Does aggregating demand into a daily average make sense here?
2. How would you extend this to multi-period optimization?
3. Any suggestions on improving the optimization layer?

The version is basically a 0.0.1, but every feedback is well accepted. Thanks!

Repo: link

Thanks!