StockPulse

StockPulse is a local-first SCM demo that turns sales and clickstream signals into demand forecasts, safety stock, reorder points, EOQ, and service-level decisions.

Run the project

Use one file to generate the data, run the analysis, and open the dashboard on localhost:

python start_local.py

This creates a local dashboard at http://localhost:8501.

If you only want the data pipeline and CSV outputs:

python run_project.py --mode both

Outputs

• data/sales_data.csv
• data/inventory_metrics.csv
• data/kpi_snapshot.csv
• predictions/predicted_adjustments_simple.csv
• predictions/predicted_adjustments_pro.csv

Main files

• start_local.py
• run_project.py
• dashboard/app.py
• src/processing/forecast_inventory.py
• src/ml/predictor_pro.py

Core formulas

• Moving average forecast
• Adjusted demand = forecast × (1 + surge index)
• Safety stock = Z × σd × √L
• Reorder point = adjusted demand × L + safety stock
• EOQ = √(2DS/H)
• Service level = 1 − stockout probability

Notes

• The local dashboard is dependency-free and runs from start_local.py.
• The Streamlit dashboard remains in the repo as an optional richer UI, but it is not required to run the project locally.