Analytics project / Operational optimization
Sourcing Optimization
Turning manual multi-warehouse sourcing decisions into faster, estimated-cost comparisons.
My role: Designed the sourcing workflow, decision rules, and Python/Streamlit application, from cost comparison to operational output.
01 / The challenge
The closest hub wasn't always the cheapest
For large project orders, coordinators had to find inventory across multiple warehouses and choose where to consolidate items before delivery. Those choices often favored proximity or fewer transfers, but neither guaranteed the lowest overall logistics cost.
We needed to compare the estimated cost of every viable staging option—including transfers between warehouses and the final shipment.
02 / The solution
From an order to a ranked sourcing plan
The tool brings together demand from one or more orders (or an item-and-quantity list) and checks available inventory across the warehouse network. For each staging option, a Python engine builds the required transfers, uses an internal shipping-cost API to estimate transfer and final-delivery costs, applies business rules, and ranks the resulting plans.
Coordinators can compare alternatives, inspect the item quantities and movements behind a plan, and export the selected configuration to Excel for the next operational steps.
Inside the application
Explore the Sourcing Matrix
Choose a view to follow the sourcing decision—from order setup to recommended hub, alternatives, cost rankings, transfer detail and carrier options.
The lowest-cost recommended plan
Hub B has the lowest estimated end-to-end cost among the hubs evaluated in this fictional example.
The recommendation combines estimated source-to-staging movements with the final shipment. See Ranking to compare every candidate.
03 / Impact and implementation
Faster preparation. Clearer tradeoffs.
After implementation of this tool, plan preparation improved from roughly one project per day to more than one per hour (8x+). The tool also provided a more consistent starting point for sourcing decisions: alternatives could be compared on estimated freight cost rather than proximity or warehouse count alone.
Built with prepared enterprise order and inventory data, Python, an internal shipping-cost calculator API, and Streamlit. AI-assisted tools helped accelerate parts of the Python development. The application provides an Excel export for operational follow-through. Modeled cost comparisons are not realized savings.
Decision Studio / Continue exploring
