Research Data Services

Have you acquired some new data science skills but need some data to practice with?

This guide points you to lists and repositories of free digital datasets. Use these to apply your new data analysis or visualization skills or to practice with a new tool.

Note: these datasets are not necessarily vetted for accuracy or consistency. If you are a researcher searching for datasets curated for research purposes please see our Find Data or Data Sharing and Repositories or the Dartmouth Libraries' many subject-area reference guides instead.

Looking to learn some new data science skills? See our list of upcoming workshops. Need some help or guide on a data science or digital humanities project? Contact us at: ResearchDataHelp@groups.dartmouth.edu.

Sample datasets used in tutorials and demo materials

• Seaborn datasets: sample datasets provided with Python's seaborn library.
• Rawgraphs datasets: sample datasets provided with Rawgraphs.io online visualization tool.
• Intro to Cultural Analytics datasets: datasets used in Melanie Walsh's book, Introduction to Cultural Analytics.

Datasets used by visualization teams and data journalists:

Try to re-create these teams' beautiful visualizations:

• Information is Beautiful
• The Pudding

Lists of practice datasets for data science

• "43 Free Datasets for Building an Irresistible Portfolio," Dataquest (2023).
• R. Nasrin Sucky, "All the Datasets You Need to Practice Data Science Skills and Make a Great Portfolio," Medium (2021).
• Y. Prakash, "26 Datasets for your Data Science Projects," Medium (2021).

• Data is Plural list: Google docs list of 1700 datasets (and growing)
• DH Resources for Project-Building: List of datasets for the Digital Humanities under the following categories:
  • Demo Corpora (ready to use texts)
  • Document / Image Collections
  • Linguistic corpora
  • Map Collections
  • Datasets (with a focus on humanities datasets)
  • Datasets in Other Disciplines (i.e. Social and Natural Sciences)
• Kaggle.com datasets (nearly 300,000 data user-submitted datasets)
  • Note: these are not vetted so if using for research you will need to verify the accuracy and consistency of the data

Datasets for Machine Learning

• UC Irvine Machine Learning Repository