The GoodsForecast Hack gathers over 50 teams and two hundred participants

The results of the first GoodsForecast hackathon have been summarized. Its task was to ensure the availability of goods on shelves that was solved using real data.

Moscow, October 24, 2023. The hackathon kicked off on Friday, October 20, and concluded by Sunday evening, October 22. During the 48 hours of the competition, the teams had to develop machine learning models and solve one of two problems of determining the availability of goods on shelves at no-sales intervals using the GoodsForecast.OSA industrial toolkit. 

The solutions presented by the teams were evaluated in terms of the quality and workability of the proposed models. The most promising developments of the finalists will undergo AV testes in real conditions, in Dixie chain stores. The following team were the best, in the opinion of the jury:

• Team 2, which was best at building a model without using the out-of-shelf probability estimate available in the data mart.
• Team 5, which was the best at building a model using the out-of-shelf probability estimate as one of attributes.

Each of the two winning teams received a cash prize of RUB 100 thousand.
The jury also mentioned the solutions of two more teams, Phyton 1 and GibData.

In total, the hackathon gathered 53 teams. It was held in an online format which enabled teams from different regions to participate in it.

The hackathon was co-organized by the “Opportunities Accelerator” of Lomonosov Moscow State University’s “Vorobyovy Gory” Innovation Scientific and Technology Center whose educational cluster GoodsForecast joined this summer..