Earlier work
Client project
Sales analytics for a retail client
A retail company
A sales dashboard pulling several sources into one place, so the team could see product and region performance without rebuilding the numbers by hand every week. Weekly reporting that used to be assembled manually now runs itself.
- Built with
- Python and pandas for the integration and cleaning, SQL against the source systems, and Power BI for the dashboard and scheduled reports.
- Result
- The client reported a 15% increase in quarterly revenue after acting on it, and a 70% reduction in the time spent producing reports by hand.
How it fits together
- Sales systems several sources
- Integrate and clean Python, pandas
- Data model SQL
- Dashboard Power BI
- Weekly report scheduled, not manual
The problem
Sales data lived in more than one system and was reconciled by hand into a weekly report. By the time the report existed it described a week that had already finished, and questions that were not on it needed another manual pass.
What I built
An integration and cleaning step in pandas that brings the sources together and resolves the differences between them, then a Power BI model on top with KPI monitoring, product performance, regional trends and customer segmentation. The weekly report became a scheduled output rather than a task.
Results
The client reported a 15% increase in quarterly revenue over the following quarter, and a 70% reduction in the time spent assembling reports. The revenue figure is theirs, measured on their numbers; what the dashboard did was make the underperforming regions visible early enough to act on.
What I would do next
Segmentation was descriptive. With the history now in one place, the same model supports forecasting per region, which is the question the team kept asking the dashboard and it could not answer.