I am working on a regression model.I need a big dataset for this purpose. I have already worked with the famous Airlines dataset(http://stat-computing.org/dataexpo/2009/) and the million song dataset (http://labrosa.ee.columbia.edu/millionsong/) .I want another dataset to gain deeper insights into the research problem. Most of the datasets which I found are not so big in UCI repository.

Can anyone suggest a source for this purpose?

  • How big does it have to be? Any other requirement, on the topic, the types of variables, etc. ? Jun 21, 2015 at 17:52
  • @PeterSmith ,I need atleast one million records.The more the better.It will be good if number of categorial variables are less.The more the continuous variables,better it will be for the purpose. Jun 21, 2015 at 20:28

1 Answer 1


At least a million records you say? I know of a data set that could fit the bill.

The 2013 ACS 5-Year PUMS has a record of 7,300,520 households (which represents 6,671,272 housing units and 629,248 group quarters units) and 14,988,864 persons. It is a nationally-representative social survey for the United States.

  • Well.thanks for sharing it ,but when I open it ,it has lot of categorial variables and it shows it to be top-coded. Jun 27, 2015 at 6:12
  • Fair enough. It is your choice to decide how useful it is to you. I only bring it to your attention.
    – Kotebiya
    Jun 27, 2015 at 13:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.