Best Data Science Competition Websites in 2026

Data science competitions are a great resource to practice and improve our data science skills. That’s why in this article, I summarize the best data science competition websites where you can find the projects you like to improve the skills that you need.

When I started my journey as a data scientist, back in 2015, the best way to learn the skills was through online courses. These courses were great at theory, but poor at offering opportunities to practice the learned skills. That’s why I turned to data science competitions.

Back then, the go-to website for finding competitions was Kaggle. But today, there are many more opportunities to find data science competitions that we like and train the skills that we need.

Some of us may be interested in practicing skills for tabular data, some for computer vision. Some of us are interested in tackling finance problems, others in helping nonprofit organizations. The good news is that with the new data science competition platforms we can find solutions that suit us all.

Let’s dive in.

Best Data Science Competition Platforms

Kaggle

Kaggle still remains the most widely used data science competition platform. It has by far the most users; it ran the most competitions in 2025, and it offered the largest total prize pool.

Kaggle offers competitions with cash prizes, public datasets for exploration and modeling, notebooks for coding and collaboration, and learning resources including tutorials and courses. It enables users to build, test, and showcase models, collaborate with others, and benchmark their skills against a worldwide community, making it a hub for both learning and professional growth in data science.

Codabench

Codabench, originally CodaLab, hosted the third-most competitions in 2025, and its user count more than doubled over the year.

Codabench is an open-source platform designed for organizing and participating in reproducible AI and machine learning competitions and benchmarks. It allows users to submit algorithms, datasets, or evaluation metrics and automatically ranks participants on leaderboards, supporting a wide range of tasks from supervised learning to complex challenges.

Codabench emphasizes transparency, reproducibility, and flexibility, making it ideal for researchers, educators, and organizations looking to evaluate models, collaborate on challenges, and advance AI research.

Zindi

Zindi is an African platform that hosts data science competitions aimed at solving real-world problems across the continent, including areas like agriculture, healthcare, and finance. It connects data scientists with impactful projects, offering opportunities to build models, collaborate with peers, and gain practical experience while tackling challenges relevant to African communities.

Zindi also provides a community-driven environment, fostering learning, networking, and skill development for both beginners and experienced practitioners in data science.

DrivenData

DrivenData focuses on solving social challenges. It connects impact-driven organizations with data scientists through competitions focused on social good, tackling challenges in areas like public health, education, and humanitarian aid.

DrivenData enables data scientists to apply their skills to real-world problems, build predictive models, and contribute to measurable positive change. It also fosters a community of socially conscious data scientists, providing opportunities for collaboration, learning, and showcasing work that has tangible societal impact.

Solafune

Solafune is a platform that hosts global data science competitions using satellite and geospatial data to address real-world challenges such as environmental monitoring, land use analysis, and climate change. It provides participants with access to rich datasets, evaluation metrics, and leaderboards, enabling data scientists and researchers to develop, test, and benchmark models while contributing to impactful solutions for pressing global issues.

MachineHack

MachineHack is a platform that hosts AI and machine learning hackathons where data scientists and developers can compete on real-world challenges using provided datasets. It offers leaderboards, evaluation metrics, and prizes, enabling participants to build, test, and benchmark models, sharpen their skills, and gain practical experience while solving meaningful problems across diverse domains.

Datacamp Competitions

DataCamp hosts simpler data science competitions where learners of all levels can practice real‑world analysis by submitting notebook‑based solutions to problems, get feedback through peer voting and judging, build a portfolio of shareable work

The skyscanner of data science competition websites

With more players into the data science competition field, there are more opportunities for us to practice our skills, engaging with projects we care about.

The downside? It becomes harder to find competitions that suit our needs. Imagine having to scan every single airline to find the cheapest or fastest flight that would take you from A to B!

Fortunately, there is a competitions aggregator that does that for you 👇

ML contests

ML contests is a platform that aggregates data science competitions from around the world.

ML contests showcases open data science competitions from the major, widely used platforms, like those mentioned in the previous section.

It also highlights competitions hosted on niche websites and by organizations focused on specialized areas, such as stock market forecasting or other domain-specific challenges.

By bringing these together, ML contests gives participants a comprehensive view of opportunities to learn, compete, and apply their skills across diverse problem domains.

Websites in foreign languages

There are 2 major players in data science competitions whose language is not English. If you speak Chinese or Japanese, these make good opportunities for you as well:

Tianchi

Tianchi is Alibaba Cloud’s global data science and AI competition platform that hosts a wide variety of challenges across fields like machine learning, big data, NLP, and industry‑specific problems, offering high‑quality datasets, cloud computing resources, and community support to help developers and researchers solve real business and research questions, build skills, and compete for prizes.

Signate

Signate is a Japan‑based data science and AI competition platform where participants can choose from a variety of challenges — from beginner practice events to prize‑awarded competitions. Users can build and submit predictive models using real datasets, compete on leaderboards, and even connect with potential recruiters or specialized industry initiatives.

Like this, Signate provides a practical way to sharpen skills and engage with the data science community.

Data Science competitions in 2025

In its latest report, ML contests analyzed the performance of most machine learning competitions in 2025. This is what they found:

• Kaggle continues to be the most widely used platform, with CodaBench and Zindi coming close as English speaking platforms.
• 💡 Python and PyTorch still dominate winning toolkits.
• Pandas still rules. Polars comes second with a lot of distance.
• GBMS are still the winners for tabular data.
• Transformers overtake CNNs in vision tasks, reshaping traditional approaches in computer vision competitions.
• Decoder models dominate NLP, especially Qwen models leading language competition wins.

Check out all the details in this article.

Final thoughts

When I was learning data science, practicing my skills using public datasets and data science competitions was the best and only way for me to be confronted with those challenges that I would find later on in my first job.

This practice gave me the confidence and, above all, the tools to first answer interview questions and second tackle real-world problems once I got the job.

If there is one piece of advice that I’d give, it’s practice, practice, practice.