← Case study: why a growing chain loses money
Simulated data · not a real company

Pharmacy chain dashboard

Explore the Gold layer of the data warehouse: 1,000 branches, 13 regions, 2020–2025. Pick a year or a region, or click a region bar. Numbers come from the same tables as the Power BI report and reconcile to the riyal.

Revenue (SAR)
Margin
Receipts
Average receipt (SAR)

Monthly revenue

Steady growth with no seasonality: a sign of generated data, not a real chain.

January 2025 jumps because receipts have 3.8 lines instead of 3 (data fingerprint F16).

Revenue by region

Click a bar to filter the whole page by that region.

Loss below cost by category, 2020–2025

Selling price below purchase cost. Not affected by the filters.

Branches to review

Weakest 10% of each peer group (same size band and city type) by average monthly score. A list to review, not a verdict.

BranchType · peer groupRevenue (SAR)MarginScorePeer avg

Products sold below cost

262 products were sold below their purchase cost, 2020–2025. 248 of them have a suspicious cost-to-price ratio and carry 99.8% of the loss: check the pack size and the purchase price before blaming anyone.

ProductCategoryRevenue (SAR)Loss (SAR)Loss per 1 SAR soldCost check

The report flags 314 products with a cost ratio below 0.30 or above 1.00; 248 of them actually sold below cost.

What was checked, honestly

Before naming any person, seven fraud patterns were tested. None was stronger than chance.

integrity signals tested
data gaps that block a measure
patterns where the data is not like a real chain
Show all checks
IDCheckResultVerdict

How it works: a Fabric notebook (nb_90) exports six small Gold aggregates to CSV; this page reads them with plain JavaScript, so it stays online without a Fabric capacity. Revenue 3,126,598,408.75 SAR and margin 918,000,519.14 SAR match the warehouse exactly.