3
$\begingroup$

How should we treat questions asking about obtaining particular datasets? Other SE sites (for instance, CrossValidated) explicitly consider this to be off-topic.

Personally, I don't necessarily have an opinion on the matter but I do think that--whatever we decide--it should be made explicit in the site's documentation.

One possible approach (taken by Quant SE, apparently) is to explicitly discourage questions of this nature but then provide a community wiki addressing data sources.


In addition, what about stock market data--should we consider that separately? (Since, for example, we might want to allow general economic data-related questions but (explicitly) exclude financial markets data.)

$\endgroup$
  • $\begingroup$ Just to clarify, are you asking about questions relating to obtaining a particular dataset (e.g., how do I get access to CRSP?) or asking about about questions related to finding even the appropriate dataset for a particular topic (e.g., what kind of data exists to help me measure the impact of class size on student acheivement?)? $\endgroup$ – jmbejara Nov 24 '14 at 18:11
  • $\begingroup$ @jmbejara: Either? I don't think there's anything wrong with the second question--I'm referring to something more blunt like, "Where can I find PPI data for Venezuela?". $\endgroup$ – Steve S Nov 24 '14 at 18:22
  • $\begingroup$ Agreed - there are some countries/data sets that really are very difficult to find, and others where they´re there, but not in the usual place, so knocking these off as individual answers wouldn´t hurt. $\endgroup$ – Lumi Nov 24 '14 at 18:31
  • $\begingroup$ There are currently 50 data-request questions: economics.stackexchange.com/questions/tagged/data-request Most of them would be fit for opendata.stackexchange.com $\endgroup$ – Nicolas Raoul Aug 24 '15 at 1:38
6
$\begingroup$

Allow them

I don't see a reason why we should forbid them. In contrary to questions for literature or good textbooks, at least these questions (and answers) are not opinion-based.

On the other hand, there are so many potential data questions that a community wiki will always take more time to write than just answering any question that pops up. I don't see an advantage to the wiki approach that makes up for the huge cost of writing it.

Stock markets

To the extent that data is used to work within Economics, it is fair game. That can include the stock market, Ebola, usage of umbrellas in the Netherlands and similar. Of course, the typical disclaimer holds: If you think a different StackExchange site will be more appropriate for the type of data, or you don't think that anyone here will know it anyways, there is little use of asking the question. Which does not mean it's forbidden.

$\endgroup$
  • $\begingroup$ Interesting. I was kind of thinking that data questions would largely fall into the categories of: 1.) things you could've googled or 2.) things so technical/arcane that they'd be totally unanswered. Also, what about stock market data--should that be considered fair game as well? $\endgroup$ – Steve S Nov 24 '14 at 18:27
  • $\begingroup$ Your type 1) is not a sufficient category for exclusion from StackExchange. Typically, even for simple/googleable questions, SE wants to provide an answer such that the google-search would, in the future, provide a SE answer. Of course, if its just a keyword and it is so easy to google/find that one could just look up wikipedia, thats a different story. $\endgroup$ – FooBar Nov 24 '14 at 18:30
  • 3
    $\begingroup$ And ad 2): Data sources are the #1 question that everyone who does applied work will have. I don't think they're arcane at all. $\endgroup$ – FooBar Nov 24 '14 at 18:31
  • $\begingroup$ Re: "type 1", I was thinking that they would be excluded since you're supposed to do a minimal amount of research before asking the question. $\endgroup$ – Steve S Nov 24 '14 at 18:37
  • 1
    $\begingroup$ Sure, but I don't think data is that google-able. At least me personally, I have so often tried to google data sources for specific things and just ended up with useless stuff like stupid diagrams on a subject, but not actually the original data. But perhaps thats just me. $\endgroup$ – FooBar Nov 24 '14 at 18:42
  • $\begingroup$ Re: you're second comment: I meant that the datasets in question would be arcane. (Regardless, that's just me speculating) ... Re: follow-up: That's true, too. (In fact, even when you're on a site that you know contains the data you need, it can be a pain just to track down the raw data download...) $\endgroup$ – Steve S Nov 24 '14 at 18:54
3
$\begingroup$

Disallow them

There is already a Stack Exchange site for requests for open data sources: the Open Data Stack Exchange

It's a never-ending source of poor questions.

$\endgroup$
  • 3
    $\begingroup$ Ha! What makes them so bad?? ... Note: I was just thinking, "There really should be an SE site for exactly this kind of thing--too bad it would never get off the ground...". $\endgroup$ – Steve S Nov 24 '14 at 19:56
  • 1
    $\begingroup$ OD.SE is limited to open data. And it works. $\endgroup$ – Anton Tarasenko Nov 26 '14 at 15:48
0
$\begingroup$

Allow

The appropriateness of using a data source for a particular question often depends on how the data is collected and how the variables are defined, and a beginner may fail to make a good judgment on these issues just by reading the documentation.

$\endgroup$

You must log in to answer this question.

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