The need
Two questions from the business: where should marketing campaigns be concentrated, and which strategy suits which part of the market? A uniform national campaign would dilute the budget, so the answer had to separate regions and installation sizes.
What I owned
All of it, as part of my application: the framing, the data preparation, the dashboard, the recommendation and the presentation. I also taught myself Power BI for this project.
The solution
An interactive dashboard over public data from data.gouv, built to answer the business questions and not simply describe the data.
How it was built
The story came first. I defined the business objective and the analytical questions before touching the data, then checked the figures against external references. The main obstacle was geographic data, where the same region appeared under different spellings and languages, such as Bretagne and Brittany.
Results
- Concentration. Five regions lead on volume, and the top five hold about 70% of the 127.54M kWh produced each year.
- The volume and value paradox. Small sites are 58% of projects, while large sites are only 9% of projects but 41% of production.
- Timing. After a peak around 2010 and a long decline, the market has recovered strongly since 2022, especially on medium and large sites.
The recommendation was a two-speed approach. A volume engine targets small and medium sites in the leading regions with digital lead generation, simulators and quick quotes. A value engine targets large projects with direct outreach.
A bonus proposal covered two AI agents for lead handling and customer support, with human validation built in. It set out the direction my internship then took.