Intermarché invites the community of data scientists and AI experts to participate in a challenge . Participants will have to develop and propose a predictive model to estimate the daily sales volumes achieved by the various outlets of the company for various consumer products. The winner of this competition will be able to walk away with a cheque for 3,000 euros.

Developing a predictive model

To carry out this mission, the candidates will have a learning base containing the daily sales of a selection of products. The model must be able to predict sales forecasts over a three-month horizon. From this information, the participant must develop a code in the form of a Python script or a Jupyter Notebook, allowing him to generate in a CSV file, the sales predictions.

The algorithms will then be tested by comparing them to the volumes realized over the first quarter of 2019, based on the Root Mean Squared Logarithmic Error (RMSLE) :

n is the number of expected predictions, i.e. for which an actual sale of the product was indeed observed on the same day or in the previous six days. yi is the predicted sales volume in the first brackets and the actual sales volume in the second brackets. The calculation is the square root of logarithmic sums multiplied by the inverse of the expected number of predictions. The lower the value, the more efficient the model is considered to be.

An innovative challenge

Candidates have until June 18 to propose their predictive models and have the possibility to make as many contributions as necessary to try to climb the ranking. On June 30, prizes totaling 5,700 euros will be awarded to the best programmers.

Intermarché spoke about its motivations for launching this challenge:

“Still too little democratized, we believe that this type of initiative will allow the various players to sharpen their skills while allowing us to take new proposals for approaches and methods on our business problems and thus contribute to pushing the excellence of the Data Science community in France ever further.”

The Data Factory Intermarché team

Translated from Challenge : Intermarché vous lance le défi de concevoir un modèle prédictif innovant