By registering in competition, you accept our terms & conditions
Data Ninja is a competition organized by OLX Group.
Each edition allows you to fight with real-world big data problems.
The competition is taking place in areas of data science and data engineering.
This is already the fourth edition and this time the challenge is to build a prediction model,
that will annotate items with most relevant search phrases.
You receive 5 million items labeled with relevant labels.
Your task is to create a model, that will be able that will be able to predict the most relevant labels (search phrases) for new ads.
Extract features from descriptions and attributes, find interesting relationships,
add something extra on your side and create a model, that will predict the future better than competitive models!
There is 10,000 PLN total prize pool, awesome gadgets and possibility to join our Product & Tech team for
paid internship to integrate your solution with live OLX websites!
Your University is not on the list? Add it yourself!
You can register an account only using your student e-mail address and therefore you are not allowed to use multiple accounts.
The maximum size of a team is 2 participants.
You may submit a maximum of 5 entries per day.
You are not permitted to share code or solution specific information outside your team. However it is ok to share solution details if made available to all participants on the forums.
No, you can use both English and Polish. We believe understanding of English language is no longer a problem for young people. However feel free to communicate in e-mail of web forum in language you are more comfortable with.
Please use our Web Forum.
No, during the competition we publish results, that come from validation on only 30% of testing dataset. So you cannot be sure the final results. However as long as the model is not overfitted both results should be almost the same. We will show the final results at Competition closing Gala and publish them in the platform the same day.
Use our web forum to ask a question or start discussion.
E-mail us to contact privately.