Kaggle Datasets
A large, searchable library of free public datasets shared by the Kaggle community, from sales and marketing data to sports, health and transport. Useful for practice and portfolio projects, as long as you check the licence and the quality of each dataset before relying on it.
Platform
Other
Kaggle (Google)
All
Free
What it is
Kaggle Datasets is the dataset library on Kaggle, the data science community owned by Google. Anyone can publish a dataset, and each one has a page with a description, file preview, column notes, a licence, a usability score and public notebooks where other people have explored it. You need a free Kaggle account to download files.
Why it is useful for analysts
- Variety. You will find retail sales, e-commerce orders, marketing campaigns, HR data, football matches, flight delays and much more, which makes it easy to practise on data from the industry you want to work in.
- Formats you will meet at work. CSV and Excel files of all sizes, some with several related tables you have to join.
- Examples to learn from. The notebooks on each dataset show how others cleaned and analysed it. Read them after you have tried yourself, not before.
- SQL practice. Kaggle also hosts public BigQuery datasets, and Kaggle Learn's SQL courses use them.
The catches
Because anyone can upload, quality varies a lot:
- Many popular datasets are synthetic or heavily cleaned. They are fine for learning tools, but a portfolio built only on famous, spotless datasets (the Titanic or a generic "superstore") does not show you can handle real data.
- Check the licence on the dataset page before publishing work based on it.
- Check the source. Prefer datasets that link to an original, such as a government portal, a company's open data or a research paper.
How to use it well
- Search for a domain you care about, then filter by file type and size.
- Read the description and column notes; if nobody knows where the data came from, pick another.
- Profile it before analysing: row counts, missing values, duplicates, date ranges.
- Write down one business-style question before you open a chart tool.
Is it worth your time?
Yes, as one of two or three data sources. Combine it with official open data portals and sites such as Our World in Data for at least one portfolio project with genuinely messy, real data.
Common questions
Do I need to enter competitions? No. Competitions are about machine learning; for analyst work, the datasets and notebooks are what matter.
Is it free? Yes.
Ready to start?
Opens on Kaggle (Google) in a new tab.
Stuck on something specific?
Writing only gets you so far. If you want an answer to your situation rather than the general case, book a session and we will work through it together. Sessions are free for approved Sefism members, and a few slots open each week.
Follow along
New writing, resources and project ideas land here first.