4 Data Science Competition Platforms Other Than Kaggle
Here are some lesser known alternatives to Kaggle
Kaggle is one of the most popular data science community and is well know for hosting top tier machine learning competitions with attractive prize pool. Here are 4 other fast growing communities with challenging machine learning problems that might interest you.
Zindi
Zindi is a social enterprise whose mission is to build the data science ecosystem in Africa. As such, many of Zindi’s competitions are focused on solving problems in Africa and involves African datasets. At the time of writing there are 5 active competitions with prize money ranging between USD $3000 to USD $ 10,000. Similar to Kaggle, the context, problem description, evaluation metrics and data are clearly explained. Zindi also has a dedicated forum for each competition which provides an avenue for discussion and for participants to interact with competition sponsors or the host.
Unlike Kaggle, Zindi does not provide cloud hosted notebooks and require participants to use their own compute resources for training and evaluating machine learning models. Check out this article for free cloud GPU compute resources.
Data Driven
DataDriven aims create social impact by tackling pressing challenges using data science hence many competitions are related to health, climate change, education and conservation. At the time of writing there are 2 prize competition with USD$20,000 and $500,000 prize pool. Similar to Kaggle, the context, problem description, evaluation metrics and data are clearly explained. There is also a discussion forum and leaderboard. Participants are encouraged to share their solutions on github or blogs. DataDriven does not have cloud hosted notebooks like Kaggle hence participants are required to use their own compute resources for training models.
Tian Chi
Tian Chi is a big data competition platform by Alibaba Cloud, the cloud computing subsidiary of Alibaba Group. At the time of writing there are 4 active prize competition with prize pool ranging from USD$10,000 to ￥1,000,000 (~$USD 157,000). Over the years Tian Chi has hosted some major competitions associated with academic conference such as CVPR 2020 and AAAI 2022. Similar to Kaggle, the context, problem description, evaluation metrics and data are clearly explained. There is also a discussion forum and leaderboard. While the forum are often filled with lively discussions, many of the discussion are in Chinese therefore Google Translate might come in handy.
Similar to Kaggle, Tian Chi provides cloud hosted notebooks with unlimited CPU runtime and limited GPU runtime. Participants can choose to create private notebooks or share their solutions to the pubic forum.
Machine Hack
Machine Hack is a community which provides bootcamp, mock assessments, practice and hackathon for data science and AI enthusiasts. The hackathon section hosts various machine learning challenges where participants can compete on. At the time of writing, there is 1 hackathon with prize money of INR 70,000 (~USD1000) and many other practice hackathon for beginners. Similar to Kaggle, context, data and evaluation metrics are clearly explained to the participants. A leaderboard is used to rank the participant’s solution and there is also a discussion forum for the community to interact. While Machine Hack does not host cloud notebooks like Kaggle, they allow participants to upload Jupyter notebook for sharing with the community.
Conclusion
Most of the mentioned data science competition platforms have similar functionalities to Kaggle with the exception of cloud hosted notebook. Out of the 4 alternatives to Kaggle only Tian Chi provides cloud hosted notebooks. I have written an article on how to get getting free GPU compute resources.
These are platforms that I know of with active competitions and prize pool. Feel free to recommend other platforms in the comments below.
