Taking part in ML competitions is a great way to learn and improve your skills. You’ve probably heard of Kaggle — but there are dozens of sites listing ML competitions. In 2022, the total prize pool across all competition sites was over $5m!
For a more structured comparison of various platforms and their features, go to our platform comparison page.
What are the top data science competition platforms in 2023?
Below we list the top 16 data science competition platforms in 2023 — including Kaggle, DrivenData, AIcrowd, Zindi, CodaLab, and others. Each platform tends to have its own niche. Some have only been around for a year.
Kaggle has the biggest community, DrivenData produce detailed write-ups of winning solutions, CodaLab is open-source and allows anyone to list competitions, and Zindi mostly works with companies and data scientists in Africa.
Kaggle
Kaggle was founded in 2010, and has over 10 million users. As well as listing competitions, it hosts discussion forums, notebooks, datasets, and models. Kaggle hosted 34 official competitions in 2022, as well as many “community” competitions.
What are the best Kaggle alternatives?
Kaggle is still the biggest platform, but there are many others running interesting competitions with significant prize money. Each Kaggle competitions tends to draw over 1,000 competing teams — so competing on another platform can increase your chances of winning!
DrivenData
DrivenData focuses on running competitions with social impact, and has run competitions for NASA and other organisations. Competitions are always followed up by in-depth research reports describing the solutions. It was founded in 2014, and has around 100,000 users.
AIcrowd
AIcrowd started as a research project at EPFL in 2017, and is now one of the top five competition platforms. It has hosted several official NeurIPS competitions, and has around 60,000 users.
Zindi
Zindi has a very active community, and is focused on connecting organisations with data scientists in Africa. Zindi also runs in-person hackathons and community events. It was founded in 2018, and has around 60,000 users.
CodaLab
CodaLab is an open-source competitions platform founded
in 2013, with an instance maintained by Université Paris-Saclay. Anyone can sign up and host or take part in a competition. Free CPU resources
are available for inference, and competition organisers can supplement this with their own hardware.
CodaLab had over 120,000 users as of October 2022.
Tianchi
Tianchi was founded in 2014, as part of Alibaba Cloud, and is the second-biggest global platform after Kaggle (by number of competitions and total prize money). The number of users is not public.
Signate
Signate was founded in 2014 and is the largest Japanese-language ML competition platform, with over 85,000 users.
EvalAI
EvalAI was founded in 2017, as part of CloudCV. It has around 30,000 users.
Waymo
Waymo is a self-driving car company that started as a project within Google. Waymo published the Waymo Open Dataset — intended for autonomous driving research — in 2020, and since then has been running annual competitions on its Open Dataset Challenges site using the dataset. There are usually 4 competitions every year with $22,000 in prizes each, for a total of $88k in annual prizes.
Xeek
Xeek was founded in 2022 as part of Shell’s Studio X, and has around 3,000 users. Xeek hosted 4 competitions in 2022.
DS Works
DS works was created by SberCloud in 2021. It is unclear how many users the platform has.
MOFC
The Makridakis Open Forecasting Center (MOFC) conducts forecasting research and runs leading time-series forecasting competitions.
Microprediction
Microprediction runs ongoing time-series prediction challenges. It paid out around $50,000 in 2022. There are 1,000 users in the Slack group and around 500 live autonomous algorithms making predictions.
Battlecode
MIT’s Battlecode has run an annual programming competition most years since 2003. While this competition can be tackled without ML or data science skills, it is well-suited to reinforcement learning approaches.
Topcoder
Topcoder is a competitive programming site which hosts some data science competitions.
Hugging Face Competitions
Hugging Face launched its 🤗 Competitions platform in February 2023.
ML Contests
ML Contests is a directory of ongoing machine learning and data science competitions, and lists competitions from all the platforms mentioned on this page. As well as listing competitions, ML Contests also publishes research on competitive machine learning.
Other platforms
Some other sites which host different forms of competitive ML:
- Numerai is a crowd-sourced quant fund that has paid out over $44m to data scientists since inception.
- crunchdao is a crowd-sourced quant fund, sharing returns across over 2,000 data scientists.
- OpenML is an open platform for sharing datasets, algorithms, and experiments. It hosts over 5,000 datasets and 24 benchmark suites, allowing researchers to review the performance of different algorithms in an open and reproducible way.
- CodaBench is a new platform, currently in beta, which supports CodaLab-style competitions with prizes and deadlines as well as OpenML-style ongoing benchmarks.
The State of Competitive ML in 2022
In our report on the state of competitive machine learning in 2022, we review the 200+ machine learning competitions that took place in 2022 and give an overview of winning solutions.