The problem

Procurement signals arrive in different formats, at different times, and with different definitions. A chart alone is not enough if nobody can tell whether its data is current or trustworthy.

The approach

This project treats data provenance as part of the product. Collection, normalization, source health, and presentation are designed as one workflow so users can see both the market signal and the quality of the evidence behind it.

From monitoring to decision support

The next layer is a read-only Purchase Scenario Simulator. It lets a buyer test proposed receipt months and quantities across multiple materials, then compare the scenario with the baseline plan before any operational record is changed.

For each material and month, the calculation combines opening stock, confirmed inbound, simulated receipts, forecast demand, safety stock, and optional warehouse capacity. The result surfaces projected closing stock, the first stock-out month, capacity warnings, and the scenario purchase value.

The interface supports filtering to baseline-risk materials, adding or removing multiple purchase lines, and reviewing monthly results per material.

Safety and data boundaries

Verified implementation

The multi-material calculation engine, API validation, and interactive production prototype have automated coverage. Type checking and 13 Cloudflare-focused tests passed during the recorded release, and the deployed multi-material API flow was verified without writing operational data.

Open the public Purchase Scenario Simulator →

What it explores

The public version intentionally omits production screenshots and internal implementation details.

← All projects