Whenever you hear the term AI, you must think about the data behind it.

In this post, I am sharing a collection of open source data sets available, to actually train the Machine Learning model to perform various actions.

A data set is a collection of data. In ML projects, we need a training data set.

1. Xray-Images

• https://ceb.nlm.nih.gov/repositories/tuberculosis-chest-x-ray-image-data-sets/
• https://www.kaggle.com/nih-chest-xrays
• http://academictorrents.com/details/557481faacd824c83fbf57dcf7b6da9383b3235a
• https://nihcc.app.box.com/v/ChestXray-NIHCC

2. US Government

• Data.gov
• NOAA – ncfc.noaa.gov/cdo-web (motions, inflation, environmental data)
• US Census Data – census.gov/data.html (demographics)
• Bls.gov/data – (employment/un-employment, product categories)

3. UK Government

• UK Dataservice – www.ukdataservice.ac.uk (census data)
• WorldBank – datacatalog.worldbank.org (census, demographics, geographic, health, income, GDP)
• IMF imf.org/en/Data (economic, currency, finance, commodities)
• OpenData.go.ke
• Data.world

Find your Fun Application ideas using these dataset:

• Kaggle.com/datasets (variety)
• snap.stanford.edu/data/web-Amazon.html (35 Million product reviews)
• Group lens.org/datasets/movielens (20M MOVIE ratings)
  Yelp.com/dataset
• IMDB – ai.stanford.edu/~amaas/data/sentiment/ (25M Movie ratings)
• Twitter Sentiments – help.sentiment140.com/for-students (160k Tweets)
• AirBnb – insiderairbnb.com/get-the-data.html
• UCI ML Datasets – mar.cs.umass.edu/ml
• EMAIL dataset – cs.cmu.edu/~enron/ (500k Emails)
• SpamBase – archive.ics.uci.edu/ml/datasets/Spambase (emails)
• reddit.com/r/datasets/comments/1uyd0t/200000_jeopardy_questions_in_a_json_file/ (200K questions and answers)
• Gutenberg EBOOKS – Gutenberg.org/wiki/Gutenberg:Offline_Catalogs (LARGE collection of ebooks)

Training Images using Natural Language Processing:

1. ImageNet – httpimage-net.org (14M images).
2. Google – ai.googleblog.com/2016/09/introducing-open-images-dataset.html (9M images URLs with labels)
3. Microsoft Coco – cocodataset.org (330k Images, mostly labelled)
4. Stanford Dogs – vision.stanford.edu/aditya86/ImageNetDogs (120 dog breeds, 20K images)

   Please comment below if you are pridicting something out of it.