25+ Best Websites to Find Datasets for Data Science Projects (2026 Edition)

Data is the lifeblood of data science. Whether you‘re just starting to learn data science and machine learning or you‘re an experienced practitioner, working with real-world datasets is essential. Datasets allow you to practice data manipulation, analysis, and modeling techniques. They provide opportunities to tackle novel challenges and gain experience working with different data types and structures. And they are the proving ground for building data science portfolios and advancing your skills.

But where can you find high-quality datasets to fuel your data science journey? In this article, I‘ve curated a list of over 25 of the best websites to obtain datasets, updated for 2024. These resources span everything from general-purpose repositories to specialized datasets across domains like healthcare, physical sciences, business, and more.

My goal is to equip you with an arsenal of go-to places for datasets so that you always have interesting data to analyze and learn from. I‘ll start with some of the most popular, broad repositories and then highlight websites to find data for more specific applications. Let‘s dive in!

General-Purpose Dataset Repositories

First up are the big names in dataset hosting that cover a wide range of topics and fields. If you‘re looking for datasets and aren‘t sure where to start, these websites are a great first stop.

1. Kaggle Datasets

Kaggle, a subsidiary of Google, is an online community of data scientists and machine learning practitioners. In addition to hosting data science competitions, Kaggle has an extensive repository of datasets uploaded and maintained by its community.

With over 50,000 public datasets as of 2024, Kaggle Datasets covers a diverse array of topics including health, science, sports, and more. You can search for datasets using keywords, filter based on dataset size, or sort by trending or most votes. Kaggle Datasets also supports discussion threads and kernel notebooks linked to each dataset.

2. UCI Machine Learning Repository

The UCI Machine Learning Repository is a collection of databases, domain theories, and data generators that are used by the machine learning community for empirical analysis of algorithms. The archive was created in 1987 by David Aha and fellow graduate students at UC Irvine and continues to be widely used in academia and industry.

As of 2024, the repository contains over 600 datasets. While the total number of datasets is fewer than Kaggle or some other resources, the UCI ML Repository focuses on high-quality and benchmark datasets that are commonly used for machine learning research. Datasets span categories including computer vision, healthcare, physics, and more.

3. data.world

data.world is a cloud-native data catalog that aims to be a "social network for data people." It provides an open data community for users to share and collaborate on datasets.

A strength of data.world is that it provides tools for users to version, document, and query datasets directly on the platform. Many of the datasets include detailed metadata, usage examples, and discussion threads to facilitate understanding.

4. Google Dataset Search

Google Dataset Search is a search engine for datasets. It leverages dataset metadata and content to index datasets hosted in thousands of repositories across the web. The metadata uses an open standard called schema.org.

To use Google Dataset Search, simply enter keywords like you would for a general Google search. The results aim to show the most relevant datasets along with detailed metadata like download location, license, and format. Because anyone can publish open data with the right metadata tags, the results may be hit or miss compared to curated repositories. However, the breadth of indexed datasets is enormous.

5. Hugging Face Datasets

Hugging Face, known for its Transformers library for natural language processing, launched the Datasets library in 2021 to improve access to NLP datasets. It provides a unified hub of over 2,500 text-based datasets with a standardized API for loading and working with the data.

The Datasets library covers NLP tasks like text classification, question answering, and language modeling. It aims to handle data processing steps like parsing, tokenization, and featurization. While narrowly focused on NLP, it‘s a great resource for practicing applied language AI.

Other notable general-purpose dataset repositories include:

• DataHub.io: A community platform for sharing high-quality datasets in a unified format
• Amazon AWS Open Data Registry: A collection of publicly available datasets hosted on Amazon Web Services
• Microsoft Research Open Data: A repository of datasets contributed by Microsoft Research for use in research and education
• European Union Open Data Portal: A catalog of open data published by EU institutions and agencies
• OpenML Datasets: A collection of benchmark datasets for evaluating machine learning algorithms

Government & Public Datasets

Governments and public institutions are massive producers and collectors of data. In recent years, there‘s been a push to make much of this data open and accessible to the public. Let‘s look at some of the best websites to find open government data.

11. data.gov

Data.gov is the United States federal government‘s open data site, providing access to over 335,000 datasets as of 2024. The data covers a wide range of sectors including agriculture, education, energy, finance, and health.

You can search for datasets by keyword, browse by topic, or filter by location, file format, organization type, and more. While some of the datasets are only relevant to government use cases, many can fuel interesting data science projects with a bit of creativity.

12. data.gov.uk

Similar to data.gov, Data.gov.uk provides a catalog of open data published by central government, local authorities and public bodies in the UK. Over 47,000 datasets spanning environment, mapping, government spending, society, and other topics are available.

13. World Bank Open Data

The World Bank is a global partnership working to end poverty and build shared prosperity in developing countries. As part of this mission, the World Bank publishes a trove of economic and development data.

Datasets include indicators on population demographics, GDP, poverty rates, and much more for countries worldwide from 1960 to present. The data is accessible both via bulk download and through an API.

14. UN Data

The United Nations provides public access to its statistical databases through the UNdata website. It contains over 60 million records covering a wide array of socio-economic and environmental indicators such as education, energy, environment, food and agriculture, and more.

15. OECD Data

The Organisation for Economic Co-operation and Development (OECD) is an intergovernmental organization with 38 member countries that works on establishing evidence-based international standards. OECD Data provides access to datasets on economics, social issues, healthcare, inequality, and more that underlie policy recommendations.

Other government and public data resources include:

• Eurostat: Statistical office of the European Union, providing high quality statistics and data on Europe
• Global Health Observatory data repository: World Health Organization‘s data collections on global health priorities and emergencies
• NASA Open Data Portal: Datasets on aerospace, sciences, and engineering collected by NASA missions
• NOAA National Centers for Environmental Information: Datasets on weather, climate, geophysics, and more collected by the National Oceanic and Atmospheric Administration
• CERN Open Data Portal: Datasets from particle physics experiments conducted at CERN, the European Organization for Nuclear Research

Datasets for Specific Domains

While the general-purpose and government data repositories offer datasets across many fields, sometimes it‘s helpful to look in more targeted places when you have a specific domain in mind. Here are a few of my favorite resources for datasets in business, audio, film, and language.

21. Yelp Open Dataset

Yelp, the popular platform for crowd-sourced reviews about businesses, publishes a subset of its data for use in research and education. The dataset includes millions of user reviews, business attribute, and user information. It‘s a great dataset for practicing sentiment analysis, building recommender systems, and analyzing trends in local businesses.

22. Million Song Dataset

The Million Song Dataset is a freely available collection of audio features and metadata for a million contemporary popular music tracks. The dataset was created in 2011 by The Echo Nest and LabROSA and is still widely used today in music information retrieval.

For each track, the dataset includes audio features such as pitch, timbre, and loudness as well as metadata like artist, album, and year. It‘s a go-to dataset for music recommendation engines and audio machine learning.

23. MovieLens

MovieLens is a web-based research recommender system that debuted in 1997. The MovieLens datasets are a series of datasets collected over the years containing anonymous movie ratings from the MovieLens community.

The datasets are benchmarks in the recommender systems research community, enabling the study of collaborative filtering algorithms. They are also popular for introductory data science projects in recommendation.

24. Linguistic Data Consortium

The Linguistic Data Consortium (LDC) is an open consortium of universities, libraries, corporations and government research laboratories that creates, collects, and distributes speech and text databases. LDC hosts datasets across over 100 languages to support language-related education, research and technology development.

Accessing some datasets requires membership in LDC, but many are freely available for research use. The datasets are widely used for building and evaluating natural language processing models.

25. Awesome Public Datasets collection on GitHub

To round out the list, I want to mention the Awesome Public Datasets repository on GitHub. This community-maintained collection provides a categorized list of datasets organized by domain, including agriculture, biology, computer networks, economics, education, and many more.

For each dataset, the repository provides a brief description and a link to access it. While light on details, it‘s a great jumping off point to discover new datasets across a variety of fields that you may not find on the other major repositories.

Conclusion

We covered a lot of ground in this whirlwind tour of over 25 top websites to find datasets for data science projects. Whether you‘re looking for general datasets to practice core data science skills or specialized datasets to tackle domain-specific challenges, these websites have you covered.

But a dataset is only as valuable as what you do with it. So I encourage you to explore these resources, find a dataset that piques your interest, and dive in! Take the opportunity to practice data cleaning, exploratory analysis, visualization, and modeling. Share your insights and results with the community.

I‘d love to hear what datasets you found most interesting and what you were able to create with them. If you know of any other excellent dataset resources, please share them in the comments to help the community.

Now, pick a dataset and let‘s get to work! And if you found this article helpful, consider sharing it with other aspiring data scientists. Let‘s learn and grow together.