Data challenges

Challenges 2024

Challenges are proposed by public services, companies or scientific laboratories, and are based on real-life problems. Participants submit the results of their classification or prediction algorithms, which are then put into competition via the website. The challenges are integrated into Pr Stéphane Mallat's course at the Collège de France, and are offered in many data science courses in France and the French-speaking world.

The 2024 edition has now been launched with seven new challenges on a variety of themes, organized in partnership with the École normale supérieure and the Institut Louis Bachelier.

Challenges

Sequentially predict career development

Presented by HrFlow (le 24/01/2024)
Presenter : Mouhidine Seiv
The aim is to use sequential decision-making methods to predict the stage at which an employee is likely to stop progressing through positions in the company hierarchy, all based on employee and company data.

Learning radiological and oncological anatomy with few shots learning

Presented by Raidium (01/17/2024)
Presenter : Corentin Dancette
The aim is to segment structures on CT-Scan images using their shape, but without exhaustive annotations. One of the difficulties lies in the fact that only certain training images are segmented.

Anticipate crowds at SNCF-Transilien stations

Presented by Transilien-SNCF (on 17/01/2024)
Presenter : Rémi Coulaud
The aim of this challenge is to predict the number of validations per day and per station over the medium-to-long term. This is a time series forecasting problem, with the complexity arising from the multiplicity of series. This challenge will enable the company to offer more appropriate services and improve operating performance.

High-frequency market data : can you identify the  stock?

Presented by CFM (on 17/01/2024)
Presenter : Stephen Hardiman
The aim of this challenge is to try to identify, from a sequence of stock market data, which is the corresponding stock, all using data extracted from the order book. A great deal of information is provided to help participants find clues to the corresponding stock.

Corrosion detection in steel pipes

Presented by SLB (24/01/2024)
Presenter : Ana Escobar
Using excerpts from topographic images of steel pipes, the aim will be to successfully segment new images to identify possible traces of corrosion.

Soccer : who will win ?

Presented by QRT (24/01/2024)
Presenter : Wissem Braham
The challenge is to predict the outcome of soccer matches. Using real historical data extracted from numerous leagues, at both match and player level, the aim will be to build a predictive model that can work for any league, level and geographical location.

Electricity price prediction

Presented by Elmy (01/17/2024)
Presenter : Anthony Galtier
The exercise consists of supervised modeling of the electricity price differential between the Intraday market (on the same day) and the SPOT market (on the previous day). Above all, it will be important to predict whether the Intraday price will be higher or lower than the SPOT price.