More than half of the questions on OpenData are questions, so I think it is time to think a bit more about:

  1. How would the ideal question look like?
  2. What are the minimum things that a question must have?

1 Answer 1


The ideal data request

  1. Data: Explain exactly what data you are looking for. Put a Wikipedia link under each ambiguous word or abbreviation, as abbreviations can have a different meaning in different fields.
  2. Context: What are you actually trying to achieve, what is your final goal that the data will help you get done?
  3. Region: Not everybody is in the USA. Say what country/region your question applies to, if applicable. Some questions do not need a region, for instance questions about chemistry or astronomy. If you want data that covers the whole globe, say it explicitly as well.
  4. License: Say what licenses are acceptable, and whether you are ready to invest money on it or not.
  5. Format: Say what numeric units and data/file format you want, if you have a preference. Say whether screen-scraping answers are acceptable or not.
  6. Authority: What kind of organization do you want the data to come from? Government-issued data only? Peer-reviewed data only? Is crowdsourced data OK?
  7. Requirements: List all other requirements you have.
  8. Non-answers: If you have tried a few obvious candidates before (or while) asking the question but they don't fit, then explain why. For instance, if you are looking for an encyclopedia but Wikipedia is not a valid answer, then explain why. This will provide more clues to what you are looking for.

See a full example question.

The minimal data request

  1. Data: Explain exactly what data you are looking for.
  2. Context: Optional, but it might prevent us from helping you or making the question more useful. Questions with context are also often more understandable.
  3. Region: Strongly recommended. If not specified, be prepared to accept answers that only apply to Pakistan or Botswana.
  4. License: Optional, we will assume you want open data.
  5. Format: Optional, but then be prepared to accept answers in esoteric formats.
  6. Authority: Optional, we will assume that you are ready to accept answers from potentially unreliable sources if no other source seems to exist.
  7. Requirements: Optional.
  8. Non-answers: Strongly recommended. At least show that you have searched for a solution and explain what is wrong with the solutions you have found so far.

Anything I forgot?

  • I think this is good (and now I understand more about what you're doing on the main site.) If we're going to lay it out like this, I would also wonder if there's a respectful but decisive way to ask people not to merely post "wishes" -- many of the data requests are for things which will never be open data, or which would require Wikipedia/OSM class efforts to gather. Mar 24, 2015 at 12:25
  • @JoeGermuska: I believe that even difficult-to-fulfil data requests are interesting. Imagine that OSM does not exist, and someone asks for it: Thanks to search engines, OpenData will be the gathering point of people looking for this data, and together they might start the collaboration. Quite a few new datasets were actually created as a direct consequence of the question being posted here, and hopefully it is only the beginning :-) Mar 24, 2015 at 14:45
  • @NicolasRaoul Sometimes, yes. But I don't think that's the same as saying that all data requests are equally valuable, and I think there's a real problem with having the unanswered question queue cluttered with essentially unanswerable questions. This meta-question is a great start towards giving guidance for asking better questions. Mar 24, 2015 at 16:20
  • @JoeGermuska: Look at softwarerecs.stackexchange.com 58% of unanswered question is perfectly livable, the community over there does not consider it as a problem: it is natural that most software just does not exist yet. Mar 25, 2015 at 3:00
  • Not looking for the last word, but just because that site exists does not persuade me that ours should work that way. I would find that site unuseful for the same reasons I think that certain kinds of data request questions are needless clutter on the Open Data SE. Mar 25, 2015 at 16:48
  • should we make this information available through the FAQ? Would be very useful to point new user to this information before they ask their questions.
    – magdmartin
    Apr 29, 2015 at 12:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .