Datasets Collection

The last modifications of this post were around 3 years ago, some information may be outdated!

👉 Note: Resources for DS & ML & DL.

Articles

• Elite Data Science -- Datasets for Data Science and Machine Learning

Create artificial dataset

• sklearn dataset module: from sklearn import datasets. This contains also some popular reference datasets.

Source of datasets

• awesome-public-datasets — A topic-centric list of HQ open datasets.
• Built-in datasets in Scikit-Learn.
• BuzzFeedNews/everything — data from BuzzFeed.
• COCO -- Common Objects in Context.
• Data Hub Datasets collection — high quality data and datasets organized by topic.
• data.gov — a large dataset aggregator and the home of the US Government’s open data.
• data.world -- The Cloud-Native Data Catalog.
• FiveThirtyEight — hard data and statistical analysis to tell stories about politics, sports, societal matters and more.
• Google Dataset Search.
• Google Trends Datastore
• Google AI Datasets — In order to contribute to the broader research community, Google periodically releases data of interest to researchers in a wide range of computer science disciplines.
• Kaggle Datasets.
• NLP-progress.
• Open Images V6
• Quandl — your perfect choice for testing your machine learning algorithms and don’t waste your time on cleaning data.
• r/datasets.
• Stanford Large Network Dataset Collection.
• UCI
• TensorFlow Datasets
• The Yahoo Webscope Program
• torchvision.datasets
• WHU-RS Datasets -- Dataset Collection by Group of Photogrammetry and Computer Vision (GPCV) at Wuhan University.

Specific Datasets

• COCO Dataset -- a large-scale object detection, segmentation, and captioning dataset.
• Dataset samples from Machine Learning Mastery.
• Fruit-Images-Dataset — A dataset of images containing fruits and vegetables.
• google-landmark -- Dataset with 5 million images depicting human-made and natural landmarks spanning 200 thousand classes.
• ImageNet -- ImageNet is an image database organized according to the WordNet hierarchy.
• Insight - BBC News Datasets
• Large-scale CelebFaces Attributes (CelebA) Dataset
• Large Movie Review Dataset (IMDB)
• MIT Places Database for Scene Recognition.
• Sarcasm detection dataset.
• UEA & UCR Time Series Classification Repository
• WordNet -- A Lexical Database for English.

Vietnamese

👉 Note: Resources for DS & ML & DL.

• IWSLT'15 English-Vietnamese data (small from Stanford).
• NLP-progress - Vietnamese
• PhoBERT -- Pre-trained language models for Vietnamese.
• PhoW2V (2020): Pre-trained Word2Vec syllable- and word-level embeddings for Vietnamese.
• ViText2SQL (EMNLP 2020 Findings): A dataset for Vietnamese Text2SQL semantic parsing.
• VnCoreNLP (NAACL 2018): A Vietnamese NLP pipeline of word (and sentence) segmentation, POS tagging, named entity recognition and dependency parsing.

Sample datasets

• Iris flower dataset (from sklearn.datasets import load_iris).
• Labeled Faces in the Wild Home (from sklearn.datasets import fetch_lfw_people).
• pydatafaker -- A python package to create fake data with relationships between tables.
• The digits dataset (sklearn.datasets.load_digits).

Tools

• TimeSynth -- A Multipurpose Library for Synthetic Time Series Generation in Python.