A lot of people might already know about these but here is a list of well known publicly available datasets you can use to play around with, from various sources.
I hope this helps you get started with your Machine Learning journey.
Amazon provides 101 different types of datasets, raging from terrain tiles, Bird occurences to space imagery.
Almost 500 different datasets in many different domains (medicine, physics, finance)
This is Google for Dataset search. No jokes. Literally.
Of course Microsoft also has their own repository of datasets. They are broken down by domain:
Biology, Physics, Social, Sciences, Information Science, Computer Science, Engineering, Healthcare, Environmental Science, Earth Science and Mathematics.
This is a repository focusing only on image datasets: Artwork from the MET, fashion related datasets, you can find it all
There are a lot (more than 15k) datasets available at Kaggle for you to play with. Below is a small list divided by problem category
- Indian Liver Patient Records
- Synthetic Financial Data for Fraud Detection
- Business and Industry Reports
- Can You Predict Product Backorders?
- Exoplanet Hunting in Deep Space
- Adult Census Income
- The Enron Email Dataset
- Ubuntu Dialogue Corpus
- Old Newspapers: A cleaned subset of HC Corpora newspapers
- Speech Accent Archive
- Blog Authorship Corpus
Time Series Analysis
Mapping and Prediction
- Seattle Police Department 911 Incident Response
- Baltimore 911 Calls
- Crimes in Chicago
- Philadelphia Crime Data
- London Crime