Case Studies

Data analysis, Power BI dashboards and automation, built end to end.

Each case study takes a common business problem and solves it with Python or Power BI, checks every result with automated tests or an independent reference, and ends with a ready-to-share result. All scenarios and data are fictitious.

Sales report with KPI cards and a monthly revenue chart

Case 01

Automated Sales Report

Brightleaf Supplies, a fictitious home-and-office goods retailer with four branches

Problem: Monthly sales reports built by hand in spreadsheets are slow, error-prone and inconsistent from one month to the next.
Solution: One command turns a sales file into a self-contained HTML report with KPI cards, charts by month, branch and category, and a top-products ranking.
Result: The full report is rebuilt in about a second, with identical numbers for identical input.

$90,267revenue analysed
500orders processed
~1 sto rebuild the report
19automated tests

Python
pandas
Matplotlib
pytest

Open live report View code

The three Power BI report pages: sales overview, branches and products, and inventory

Case 02

Sales & Inventory Dashboard (Power BI)

The same fictitious retailer, two years of sales and month-end stock across four branches and twelve products

Problem: Sales, margin and stock live in separate spreadsheets, so managers cannot see whether growth comes with healthy margins or where products run out.
Solution: A star-schema Power BI model with 20 DAX measures (time intelligence, semi-additive stock) and three report pages drawn with SVG measures in native visuals.
Result: Every number in the report matches an independent Python reference, from yearly revenue down to monthly stockout rates.

+12.9%revenue vs last year
46.3%gross margin
0.7%stockout rate
20DAX measures

Power BI
DAX
Power Query
Python

View code

What every case study includes

Business framing

Problem, solution and measurable result, in plain language.

Working code

Runs locally in a few commands, with pinned dependencies.

Automated tests

Core logic covered by tests that run before every change.

Sample output

Reports and screenshots you can explore without installing anything.

Data and privacy

All companies, people, products and figures are fictitious and generated by code with a fixed random seed.
The code and this site do not call external services, use trackers or collect any data.