100% пройдено

Competition is over

Problem description

Would you like to solve classic machine learning competitions once and for all?
Participants are invited to build their AutoML (automatic machine learning) systems on the first of a kind banking dataset of datasets: transactions, time-series as well as classic table data from real banking operations.

 Dataset  Github

Submission format

Participants have to submit their solutions in the container format. Each solution has to provide a meta-algorithm that covers the full pipeline of machine learning model building: from data pre-processing up till hyper-parameter tuning and model selection to further use the best model for target variable inference.

 Evaluation  Baseline


Total prize fund is 3 000 000 ₽!
The 1st place gets 1 000 000 ₽, 2nd place 500 000 ₽, 3rd place 300 000 ₽. 4th and 5th each get 200 000 ₽. 6th till 10th places each get 100 000 ₽. Also top 3 public solutions shared by participants on github will be awarded with 100 000 ₽ by the competition jury.

 Rules  Finale


How to participate in «Data Science Contest»?

Sign up for this competition with registration form. Develop your solution. Submit your solution and see how it ranks among others. Solutions can be submitted again.

What are the prizes?

First place - 1 000 000 RUB, second place 500 000 RUB, third place 300 000 RUB, fourth and fifth places 200 000 RUB, sixth to tenth places 100 000 RUB. Moreover top three public solutions will be awarded 100 000 RUB.

Is participation free?

Yes. Registration and participation are free.

Are people from other countries eligible to participate?

Yes they are eligible.

Is this competition solo or teams are allowed?

Participants are allowed to team up. Each team will have no more than four players.

When does the registration begin?

«Data Science Contest» runs from September 19 to November 3. Registration and solution submission will be available until 3rd November 23:59:59 (UTC+3) incl.

Could I join some time later?

Yes, solution submission will be available until 3rd November 23:59:59 (UTC+3) incl.

Do participants have to choose their final submits?

Yes. Every participant should choose up to 2 submits which will be scored for final evaluation. The best score of these 2 submits will be the final result of a participant.

When can I choose these final submits?

Choosing final submits will be available for participants from 19th September to 4th November.

What are the solution evaluation criteria?

Solutions will be scored automatically by comparison between predicted and true labels (known only to the organisers) after running models on hidden test data. Leaderboard is evaluated online and updated in real time.

When will the winners be selected?

The final evaluation of submissions chosen by participants on the hidden private test data will take place on November 4 at 12:00 (UTC+3). Winners will be selected and announced on this site until November 4th 23:59 (UTC+3)

Will there be any award ceremony?

Yes. The award ceremony will take place on November 10 in Moscow during «Data Science Day» conference. Winners of the public solutions special nominations will also be announced and awarded on «Data Science Day».

Who are eligible to participate in «Data Science Contest»?

All participants having reached the age of 18 who agreed with "Rules" and built a solution according to the description.

Does «Data Science Contest» have participation constrains?

Participation is not allowed for those who have direct or indirect relation to competition preparation of tasks and data by the organisers. Participants under these constraints who agreed with "Rules" could submit solutions but are not eligible for monetary prizes.

Are Sberbank company group employees allowed to participate in this competition?

Sberbank group employees are allowed to participate, but are not eligible for monetary prizes.

How will the best public solutions will be selected?

Best public solutions that were published on GitHub will be selected by the committee comprised of competition organisers and external experts.