Every analysis, start to finish
9 projects across Python, SQL and data visualisation. Each one starts from a business question and ends with something a team can act on.
- Python
- scikit-learn
- Pandas
Spotting counterfeit banknotes from six measurements
Comparing four machine-learning models to tell genuine euro banknotes from fakes using their geometry alone, then shipping the best one as a ready-to-use script.
Read the case study →
- Tableau
- Python
- Pandas
Which cities are worth investing in on Airbnb?
A three-page Tableau dashboard that takes an investor from the global market down to a single neighbourhood, built on 478,000 cleaned listings across 16 countries.
Read the case study →
- Python
- scikit-learn
- Pandas
Where should a French poultry brand export first?
An international market study of 164 countries: 15 PESTEL indicators, a principal component analysis and a clustering that turn public data into a three-stage export roadmap.
Read the case study →
- SQL
- dbt
- Snowflake
Who still enrols in the Data path?
A dbt and Snowflake pipeline that cleans, pseudonymises and enriches four years of OpenClassrooms enrolments with INSEE data, then reads how the student profile is changing.
Read the case study →
- Power BI
- DAX
- Power Query
Is a 7.3% budget overrun under control?
A Power BI dashboard that tracks 104 IT and marketing projects for an oral care company in 52 countries on costs, delays and deliverables, for three levels of directors, and what a second look at the data changed.
Read the case study →
- Python
- Pandas
- SciPy
What really drives sales at an online bookshop?
Two years of sales and 680,000 purchases: revenue trends, catalogue tops and flops, customer concentration, and five statistical tests on what a customer profile really changes.
Read the case study →
- SQL
- SQLite
- Excel
How did the French property market hold up in early 2020?
A relational database built from public property sales, INSEE and geographic data, then twelve SQL queries on volumes, prices and local dynamics in the first half of 2020, first lockdown included.
Read the case study →
- Python
- Pandas
- Matplotlib
Does the world produce enough food to feed everyone?
A public-health study for the FAO on four datasets: how many people are undernourished, how many the world could feed, where its food actually goes, and what a second look at data coverage changes.
Read the case study →
- SQL
- SQLite
What drives the price of a home insurance policy?
A SQLite database of 30,335 home insurance contracts, twelve business questions in SQL, and what the joins were hiding.
Read the case study →