What are the best datasets for machine learning? After scraping the web hours after hours, we have created a great cheat sheet for high quality and diverse machine learning datasets.
Last updated
October 2, 2018
May 15, 2019
A few things to keep in mind when searching for high-quality datasets:
1.- A high-quality dataset should not be messy, because you do not want to spend a lot of time cleaning data.
2.- A high-quality dataset should not have too many rows or columns, so it is easy to work with.
3.- The cleaner the data, the better — cleaning a large dataset can be incredibly time-consuming.
4.- Your end-goal should have a question/decision to answer, which in turn can be answered with data.
If you are aware of other high-quality, public datasets, which you recommend to people for research and application of machine learning, deep learning, data science, etc. Please feel free to suggest them along with the reasons, why they should be included in the comments below or by emailing Stacy directly at sstanford@mlmemoirs.xyz.
If the reason is strong, we will analyze them and include them in this list. Also, please let us know your experience with using any of these datasets in the comments section.
Happy machine learning!
Stacy Stanford, .
, , .
: Similar to how works, Dataset Search lets you find datasets wherever they’re hosted, whether it’s a publisher’s site, a digital library, or an author’s personal web page.
: A data science site that contains a variety of externally contributed to interesting datasets. You can find all kinds of niche datasets in its , from to to .
: One of the oldest sources of datasets on the web, and a great first stop when looking for interesting datasets. Although the data sets are user-contributed and thus have varying levels of cleanliness, the vast majority are clean. You can download data directly from the UCI Machine Learning repository, without registration.
: Discover computer vision datasets by category, it allows searchable queries.
: Discover high-quality datasets thanks to the collection of Huajin Wang, CMU.
: This site makes it possible to download data from multiple US government agencies. Data can range from government budgets to school performance scores. Be warned though: much of the data requires additional research.
: Contains data on how local food choices affect diet in the US.
: A survey of the finances of school systems in the US.
: Data on chronic disease indicators in areas across the US.
: Data on educational institutions and education demographics from the US and around the world.
: The UK’s largest collection of social, economic and population data.
: A comprehensive visualization of US public data.
Contains information collected by the U.S Census Service concerning housing in the area of Boston Mass. It was obtained from the and has been used extensively throughout the literature to benchmark algorithms.
An improved dataset for landmark recognition and retrieval. This dataset contains 5M+ images of 200k+ landmarks from across the world, sourced and annotated by the iki Commons community.
: A good source for economic and financial data — useful for building models to predict economic indicators or stock prices.
: Datasets covering population demographics, a huge number of economic, and development indicators from across the world.
: The International Monetary Fund publishes data on international finances, debt rates, foreign exchange reserves, commodity prices and investments.
: Up to date information on financial markets from around the world, including stock price indexes, commodities, and foreign exchange.
Examine and analyze data on internet search activity and trending news stories around the world.
: A good source to find US macroeconomic data.
: xView is one of the largest publicly available datasets of overhead imagery. It contains images from complex scenes around the world, annotated using bounding boxes.
: A large dataset of annotated images.
: The de-facto image dataset for new algorithms, organized according to the WordNet hierarchy, in which hundreds and thousands of images depict each node of the hierarchy.
: Scene understanding with many ancillary tasks (room layout estimation, saliency prediction, etc.)
: Generic image understanding and captioning.
: 100 different objects imaged at every angle in a 360 rotation.
: Very detailed visual knowledge base with captioning of ~100K images.
: A collection of 9 million URLs to images “that have been annotated with labels spanning over 6,000 categories” under Creative Commons.
: 13,000 labeled images of human faces, for use in developing applications that involve facial recognition.
Contains 20,580 images and 120 different dog breed categories.
: A very specific dataset and very useful, as most scene recognition models are better ‘outside’. Contains 67 Indoor categories, and 15620 images.
: A slightly older dataset that features product reviews from Amazon.
reviews: An older, relatively small dataset for binary sentiment classification features 25,000 movie reviews.
: Standard sentiment dataset with sentiment annotations.
: A popular dataset, which uses 160,000 tweets with emoticons pre-removed.
: Twitter data on US airlines from February 2015, classified as positive, negative, and neutral tweets
: Question answering dataset featuring natural, multi-hop questions, with strong supervision for supporting facts to enable more explainable question answering systems.
: Email data from the senior management of Enron, organized into folders.
: Contains around 35 million reviews from Amazon spanning 18 years. Data include product and user information, ratings, and plaintext review.
: A collection of words from Google books.
: A collection of 681,288-blog posts gathered from blogger.com. Each blog contains a minimum of 200 occurrences of commonly used English words.
: The full text of Wikipedia. The dataset contains almost 1.9 billion words from more than 4 million articles. You can search by word, phrase or part of a paragraph itself.
: An annotated list of ebooks from Project Gutenberg.
: 1.3 million pairs of texts from the records of the 36th Canadian Parliament.
: Archive of more than 200,000 questions from the quiz show Jeopardy.
: Archive of more than 480,000 critic reviews (fresh or rotten).
: A dataset that consists of 5,574 English SMS spam messages
: An open dataset released by Yelp, contains more than 5 million reviews.
: A large spam email dataset, useful for spam filtering.
Currently the largest dataset for self-driving AI. Contains over 100,000 videos of over 1,100-hour driving experiences across different times of the day and weather conditions. The annotated images come from New York and San Francisco areas.
Large dataset that defines 26 different semantic items such as cars, bicycles, pedestrians, buildings, streetlights, etc.
: More than 7 hours of highway driving. Details include car’s speed, acceleration, steering angle, and GPS coordinates.
: Over 100 repetitions of the same route through Oxford, UK, captured over a period of a year. The dataset captures different combinations of weather, traffic, and pedestrians, along with long-term changes such as construction and roadworks.
: A large dataset that records urban street scenes in 50 different cities.
: This dataset is useful for perception and navigation of autonomous vehicles. The dataset skews heavily on roads found in the developed world.
: More than 10000+ traffic sign annotations from thousands of physically distinct traffic signs in the Flanders region in Belgium.
: A sample of the 1,000+ hours of multi-sensor driving datasets collected at AgeLab.
: This dataset includes traffic signs, vehicles detection, traffic lights, and trajectory patterns.
: Dataset for small traffic lights for deep learning.
: Another dataset for traffic lights. This is taken in Paris.
: Datasets for traffic lights, pedestrian and lane detection.
: Openly available dataset developed by the MIT Lab for Computational Physiology, comprising de-identified health data associated with ~40,000 critical care patients. It includes demographics, vital signs, laboratory tests, medications, and more.
