TOP 15 Public Datasets for Machine Learning in 2026

Bright Data (formerly Luminati Networks) provides ready-to-use, fresh, and structured datasets sourced from over 120 domains. With a focus on high-quality, validated data, their service allows businesses to access essential datasets without the need for building scrapers or bypassing web blocks. Their datasets are designed for businesses and researchers in industries such as marketing, real estate, AI, lead generation, and financial services. Bright Data ensures ethical data collection practices, offering daily updates and flexible subscription options to suit your specific data needs.

Key Features:

• Wide Range of Datasets: Bright Data offers access to datasets from over 120 domains including LinkedIn, Amazon, Instagram, TikTok, Zillow, and more. These datasets cover topics such as social media profiles, product listings, job postings, and real estate information.
• Clean and Validated Data: The datasets are free of duplicates and errors, ensuring you get high-quality data that is ready for analysis and modeling.
• Real-Time Data Updates: Bright Data refreshes its datasets daily, with monthly updates to ensure the most current and accurate data.
• Customizable Data: Users can filter datasets according to their needs using AI-powered tools, and access data in multiple formats like JSON, CSV, or Parquet.
• Ethical and Compliant Data Collection: Bright Data maintains 100% ethical data collection practices, complying with relevant legal standards.
• Flexible Delivery Options: Data can be delivered through various methods, including API, S3, and Webhooks, to integrate seamlessly with your existing infrastructure.
• Cost-Efficient Subscription Plans: With volume discounts, strategic bundles, and tailored subscription models, Bright Data offers competitive pricing to meet the needs of businesses of all sizes.

Bright Data's datasets are perfect for businesses in need of up-to-date, real-time information for applications like AI training, market research, lead generation, and competitive analysis. For example, real estate investors can use datasets like Zillow properties and Airbnb listings to monitor market trends, while marketing teams can leverage social media data from platforms like LinkedIn and Instagram to enhance lead generation and campaign targeting.

2. UCI Machine Learning Repository

The UCI Machine Learning Repository is one of the most comprehensive and widely-used collections of datasets for machine learning research. It has served as a valuable resource for the academic community since its inception in 1987. This repository houses datasets from various domains such as biology, finance, healthcare, physics, and more, making it a versatile tool for researchers and practitioners alike.

Key Features:

• Wide Variety of Domains: The repository includes datasets related to classification, regression, clustering, and recommendation systems.
• Community Contributions: Many datasets have been contributed by researchers worldwide, ensuring continuous updates and diversity.
• Detailed Descriptions: Each dataset comes with a detailed description of the features, problem context, and sometimes even baseline performance results, which can aid in benchmarking.
• Accessibility: Data is free to download, and the repository is easily navigable.

UCI datasets are commonly used for educational purposes and as benchmarks for testing and comparing machine learning algorithms. Some of the most famous datasets in machine learning, such as the Iris dataset and the Adult dataset, are available here. The variety of datasets also makes it a go-to source for solving real-world problems using different machine learning models.

3. Kaggle Datasets

Kaggle is well-known for its data science competitions, but it also provides a vast collection of datasets. Kaggle Datasets is a repository of high-quality datasets across numerous domains such as image recognition, natural language processing (NLP), time-series forecasting, and financial analysis. Kaggle's platform also offers a collaborative environment, where data scientists and researchers can discuss, share, and refine their work.

Key Features:

• Diverse Data: From structured datasets to unstructured data like images and text, Kaggle hosts datasets suitable for nearly every machine learning task.
• Competition Data: Many datasets are sourced from Kaggle competitions, offering a real-world challenge context.
• Public and Private Datasets: Kaggle provides both open-source and private datasets. Private datasets are often used in competitions where participants must sign up to access them.
• Community Support: Kaggle enables a collaborative environment with forums where participants can discuss the datasets, share ideas, and even share kernels (code notebooks).
• Data Exploration Tools: Kaggle offers built-in tools for data visualization and exploration, making it easy for users to get started.

The Kaggle Datasets platform is ideal for those who want to quickly dive into a machine learning project. Whether you're working on a competition or learning new techniques, Kaggle’s vast array of datasets and the associated community can help you refine your skills and gain exposure to new problems.

4. OpenML

OpenML is an open platform designed to facilitate the sharing and collaboration of datasets, machine learning models, and workflows. It not only allows users to access a wide range of datasets but also enables them to share and benchmark machine learning models. The goal of OpenML is to create an ecosystem that accelerates scientific discovery by offering a transparent and collaborative approach to data science.

Key Features:

• Dataset and Model Sharing: OpenML provides a platform for sharing not only datasets but also machine learning models, making it easier to replicate results and build on previous work.
• Benchmarking: Users can benchmark their models against public datasets and compare their results with others.
• Collaborative Environment: OpenML encourages collaboration by allowing users to contribute datasets, share experiments, and discuss methods.
• Searchable Repository: The platform offers powerful search and filtering capabilities, allowing users to easily find datasets by task type, feature, or performance.
• Integration with Popular Libraries: OpenML integrates with major machine learning libraries like scikit-learn, making it easy to load datasets and train models directly in your environment.

OpenML is perfect for data scientists who need a collaborative platform to exchange datasets and machine learning models. It’s also an excellent choice for researchers looking to validate their models or compare results across multiple approaches.

5. Microsoft Research Open Data

Microsoft Research Open Data provides a collection of high-quality, public datasets spanning domains such as healthcare, environment, economics, and social sciences. These datasets are provided by Microsoft Research, along with collaborations from universities and other institutions. The initiative is designed to foster open research and collaboration, providing researchers with valuable data for advancing the state-of-the-art in various fields.

Key Features:

• Diverse Data: The datasets range across multiple domains including environmental sciences, health research, and social data.
• Real-World Applications: Many of the datasets have been used in Microsoft’s own research, making them practical and insightful for machine learning projects.
• High-Quality Standards: Data provided by Microsoft Research is often curated and well-documented, making it easier for researchers to apply machine learning techniques.
• Collaboration: Microsoft Research Open Data supports collaboration between researchers and institutions by offering data for public use.

Microsoft Research Open Data is well-suited for academic and scientific research. It’s particularly beneficial for projects requiring high-quality, reliable datasets in areas such as healthcare and environmental studies. Its focus on open research makes it a valuable resource for teams looking to push the boundaries of data-driven science.

6. Amazon Web Services (AWS) Public Datasets

Amazon Web Services (AWS) Public Datasets offers a vast collection of data hosted in the cloud, covering fields such as biology, economics, and climate science. These datasets are available for free, and they come with the added benefit of AWS’s scalable cloud infrastructure, which allows users to process large datasets quickly and efficiently. AWS’s platform is designed for users who need access to massive datasets for data analysis or machine learning tasks.

Key Features:

• Large-Scale Data: Many AWS datasets are massive in size, making them suitable for big data analysis and machine learning tasks.
• Cloud-Optimized: The data is hosted on AWS infrastructure, allowing for seamless integration with other AWS services like S3, EC2, and SageMaker.
• Diverse Data: AWS offers datasets across various domains, including genomics, satellite imagery, and more.
• Free Access: While the datasets are free to use, AWS users can also leverage the platform's computational power for analysis, though cloud computing costs may apply for processing large datasets.
• Data Formats: AWS datasets are available in a variety of formats, making them easy to integrate with different tools and programming languages.

AWS Public Datasets is ideal for data science and machine learning practitioners who need to handle large-scale datasets. The integration with AWS services allows users to scale their analysis and perform distributed computing on big data, making it an excellent option for more resource-intensive projects.

7. ImageNet

ImageNet is one of the most renowned and widely used datasets in the field of computer vision. It contains millions of images labeled with thousands of categories, making it a powerful resource for training deep learning models, especially for image classification, object detection, and feature extraction. ImageNet was pivotal in advancing the field of deep learning and remains a benchmark dataset for evaluating model performance.

Key Features:

• Large-scale Dataset: ImageNet contains over 14 million labeled images, with more than 20,000 categories, making it one of the largest and most diverse datasets for computer vision.
• High-Quality Annotations: Images are labeled with precise categories, providing clear, high-quality annotations that are crucial for supervised learning.
• Annual Competitions: ImageNet hosts the annual ImageNet Large Scale Visual Recognition Challenge (ILSVRC), which has fostered significant advancements in computer vision, including the development of deep convolutional neural networks (CNNs).
• Wide Adoption: ImageNet has been used for a variety of tasks, including image classification, object detection, and image captioning, becoming the standard for benchmarking models.

ImageNet is primarily used for training image classification models, object detection systems, and deep learning-based computer vision systems. It is widely adopted by research labs and tech companies for building robust, high-performance visual recognition systems.

8. COCO (Common Objects in Context)

COCO is a large-scale dataset designed for tasks such as object detection, segmentation, and captioning. It is a highly detailed and challenging dataset with images labeled with over 80 object categories. COCO's diverse and rich annotations include object boundaries, key points for human poses, and image captions, making it ideal for complex computer vision tasks beyond simple classification.

Key Features:

• Comprehensive Annotations: Each image in the COCO dataset includes object labels, bounding boxes, segmentation masks, and keypoints for human poses.
• Diverse Image Sources: The dataset features a wide variety of real-world images, including crowded scenes, diverse backgrounds, and different lighting conditions, making it suitable for developing robust computer vision models.
• Large-scale: COCO includes over 300,000 images and over 2.5 million labeled instances, covering a wide range of scenes and objects.
• Multifaceted Tasks: The dataset supports multiple computer vision tasks, including image classification, object detection, segmentation, and image captioning.

COCO is commonly used for training and evaluating models on object detection, semantic segmentation, and caption generation tasks. It is particularly useful for building applications that require fine-grained recognition and spatial understanding of objects within complex scenes.

9. SEER Cancer Statistics

The SEER (Surveillance, Epidemiology, and End Results) Program provides cancer-related data collected from cancer registries in the United States. SEER's datasets contain detailed information on cancer incidence, survival, and mortality, segmented by various demographic factors such as age, race, and sex. SEER data is widely used for cancer epidemiology studies, public health research, and healthcare policy development.

Key Features:

• Cancer Statistics: SEER provides detailed statistical data on cancer incidence, survival rates, and mortality across various cancer types and patient demographics.
• Longitudinal Data: The datasets cover multiple decades, allowing for long-term studies on cancer trends, survival, and treatment outcomes.
• Demographic Segmentation: The data is segmented by age, race, gender, and geographic location, enabling detailed analysis of health disparities.
• Public Health Insights: SEER data helps inform cancer prevention strategies, early detection, and treatment plans based on statistical trends.

SEER Cancer Statistics is widely used by researchers, public health organizations, and healthcare policymakers to analyze cancer trends, identify risk factors, and assess the effectiveness of cancer treatment and prevention programs. It is also a key resource for developing predictive models for cancer diagnosis and prognosis.

10. LendingClub Loan Data

The LendingClub Loan Data provides a detailed dataset of loans issued through the LendingClub platform, which is a peer-to-peer lending service. This dataset contains information about loan attributes, borrower characteristics, and payment histories. It is widely used for analyzing credit risk, developing predictive models for loan default, and building financial models.

Key Features:

• Detailed Loan Data: The dataset includes detailed records of loans, including loan amount, interest rate, term, and borrower’s credit score.
• Repayment Data: It provides information about loan repayments, including on-time payments, late payments, and defaults.
• Large Dataset: With millions of records, this dataset provides a robust foundation for developing models that predict loan performance.
• Financial Insights: LendingClub data is ideal for performing credit risk analysis, evaluating the impact of borrower characteristics on loan performance, and analyzing the financial behavior of borrowers.

LendingClub Loan Data is frequently used by financial analysts, data scientists, and machine learning practitioners to build credit scoring models, predict loan defaults, and perform financial risk analysis. It is also valuable for anyone working on predictive analytics in the fintech sector.

11. Yelp Open Dataset

The Yelp Open Dataset contains a rich collection of user-generated reviews, business information, and ratings, making it an excellent resource for sentiment analysis, recommendation systems, and natural language processing (NLP) tasks. This dataset is designed to help researchers and developers create models that can predict ratings, classify reviews, and understand user preferences.

Key Features:

• User Reviews and Ratings: The dataset includes over 8 million reviews and ratings from users, providing a rich source of sentiment data.
• Business Information: It includes data on businesses, such as location, hours of operation, and types of services, which is valuable for building recommendation systems.
• Metadata: Yelp’s dataset includes metadata like user information (anonymized) and business categories, which can be useful for clustering, classification, and recommendation modeling.
• Sentiment and NLP: Yelp reviews provide a natural language corpus that is ideal for sentiment analysis and NLP applications.

Yelp Open Dataset is extensively used for developing recommendation systems, performing sentiment analysis, and understanding customer reviews. It is particularly valuable for applications in the hospitality, restaurant, and retail sectors, where understanding customer feedback is crucial for improving services and products.

12. IMDb Datasets

IMDb (Internet Movie Database) provides comprehensive datasets related to movies, television shows, actors, directors, and crew. These datasets include detailed information such as movie ratings, plot summaries, cast lists, and much more. IMDb datasets are widely used for building recommendation systems, performing sentiment analysis, and even studying trends in the entertainment industry.

Key Features:

• Movie and TV Show Data: Includes data on movies, TV shows, actors, directors, production companies, and genres.
• User Ratings and Reviews: IMDb datasets provide ratings from users, making it ideal for sentiment analysis and understanding public opinion on media content.
• Rich Metadata: Detailed information like movie budgets, box office revenue, production dates, and cast member roles.
• Comprehensive Coverage: Data covers not only the movies themselves but also associated elements like soundtrack, reviews, and trailers, offering a holistic view of the entertainment world.

IMDb datasets are frequently used in developing movie recommendation systems, sentiment analysis models, and even for research in media consumption trends. They're also helpful in predicting movie success and analyzing the impact of actors and directors on a film’s reception.

13. U.S. Government's Open Data

Data.gov is the U.S. government's open data platform that provides access to a vast collection of publicly available datasets from federal agencies, state and local governments, and even international organizations. The platform covers a wide array of topics such as health, education, transportation, agriculture, environment, and more. Data.gov aims to encourage transparency, innovation, and the development of data-driven applications.

Key Features:

• Wide Range of Topics: Datasets cover areas like climate, energy, economics, public safety, education, and more, offering a diverse array of information for analysis.
• Government Transparency: Data.gov provides easy access to data collected by various federal agencies, enhancing government transparency and accountability.
• Public Health and Safety: Includes important datasets related to public health (e.g., COVID-19 statistics) and disaster response, useful for social research and public policy.
• Open Access: Data is freely available to the public for use in research, development, and innovation.

Data.gov is ideal for research in public policy, economics, environmental studies, and social sciences. The platform is used by researchers, developers, and government entities to create applications, visualize trends, and support data-driven decision-making.

14. World Bank Open Data

The World Bank Open Data platform provides global development data, including economic indicators, social statistics, and environmental data. The platform contains more than 16,000 datasets on topics such as global poverty, education, health, and trade. These datasets are invaluable for policymakers, researchers, and analysts working on global development issues.

Key Features:

Global Coverage: Offers data on over 200 countries and regions, covering diverse economic, social, and environmental metrics.

Economic Indicators: Includes data on GDP, inflation, employment, and trade, making it ideal for macroeconomic analysis.

Social and Environmental Data: Provides data on topics like poverty, health, education, and environmental sustainability, essential for social research and development planning.

Time Series Data: Many datasets are presented as time series, enabling longitudinal analysis of trends over time.

World Bank Open Data is widely used for economic research, development studies, and policy analysis. It is also valuable for conducting studies on global health, poverty alleviation, environmental sustainability, and social development.

15. FEMA Disaster Data

The Federal Emergency Management Agency (FEMA) provides disaster-related datasets that include information on natural and man-made disasters in the United States. These datasets offer insights into the frequency, scale, and impact of disasters such as hurricanes, floods, wildfires, and tornadoes. FEMA's data is instrumental in disaster management, risk assessment, and response planning.

Key Features:

• Comprehensive Disaster Data: Includes data on the occurrence and aftermath of natural and man-made disasters, such as affected regions, damages, and fatalities.
• Response and Recovery Data: Provides information on FEMA’s response actions, including financial assistance and relief efforts provided to affected communities.
• Real-Time Updates: Data is frequently updated with new disaster events, making it useful for real-time analysis and decision-making.
• Geospatial Data: Many datasets come with geographic information (GIS) for mapping disaster-affected areas and planning responses.
• FEMA Disaster Data is crucial for disaster response, risk management, and developing predictive models for disaster preparedness. It is commonly used by governments, humanitarian organizations, and researchers working on emergency management, public safety, and environmental science.