5

I'm working on a project for automatically identifying and sorting web analytics incidents.

As part of testing of the software, I'm looking for time-series dataset (ideally >100K rows) with daily data broken down by some well-defined dimensions (e.g. country, marketing channel etc). I'd be better if the dataset contained more than just one metric (e.g. visits and orders).

Does anyone of you know about any such public dataset?

  • It seems the documentation ( A Week in the Life of a Browse) link is not working. Do you know what the ``Events'' mean?, i.e. what is an event ? – Eyal Gutflaish Sep 14 '17 at 18:43
  • See the alternative download site in the answer below – philshem Sep 15 '17 at 8:59
2

I publish the Google Play Store statistics of one of my apps:

http://datahub.io/dataset/google-play-statistics/

See for instance the installs per device type. It is fragmented by day, with metrics such as installs/uninstalls/upgrades.

It is less than 100K rows, though, as data is pre-aggregated.

  • Thanks @Nicolas Raoul! I think it's very close to what I need. I'll definitely check it!! – Stergios Mar 30 '15 at 10:17
  • Do you think it would be possible to have such a dataset with multiple dimensions (e.g. installs per country, device type, app version etc)?? – Stergios Mar 30 '15 at 10:28
  • @Stergios: Installs per country, per device type, per app version, and many more dimensions, are already available in my dataset, you just have to correlate the various CSV files by date :-) – Nicolas Raoul Mar 31 '15 at 1:26
  • Nicolas, I don't see how I can get, for instance, installs from DE, from '10_or_more_inch_tablets' on 2015-01-12. Is there such a dataset available? – Stergios Mar 31 '15 at 15:59
  • @Stergios: Installs from DE on 2015-01-12 are available in the country-installs.csv file. Installs on '10_or_more_inch_tablets' on 2015-01-12 are available in the device_class_installs.csv file. But unfortunately, installs per country per device class per day data is not available. – Nicolas Raoul Apr 1 '15 at 1:29
2

I'm not sure if it fits your criteria, but consider the A Week in the Life of a Browser data set, which combines browsing history with survey data.

Test duration: 7 days

Test type: Global

Firefox versions covered: Fx 3.5 and Fx 3.6, Fx4 Beta

Data submission: 527,817 test sets submitted in November 2010.

Three Tables:

  1. Table Name: 'users'

    • Main table of users and the metadata associated with each user.
  2. Table Name: 'events'

    • Main table of users and recorded events.
  3. Table name: 'survey'

    • Table of users and associated Beta Background Survey answers.

Data formats are either CSV or SQL dump.

Data is licensed by Creative Commons Attribution 3.0.


UPDATE: Alternative download site

  • Thanks @philshem but there seems to be a problem with the links on the website, so I cannot download them. – Stergios Mar 6 '15 at 16:42
  • @Stergios We have tracked down this dataset and hosted it. Does it work for you? – philshem Apr 9 '15 at 10:53
  • OK I downloaded the files. Thanks a lot, I appreciate it!! – Stergios Apr 16 '15 at 10:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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