5

I'm thinking about proposing a pharmacology site for Stack Exchange. However, pharmacology is a tag on Biology.SE, among other related tags, and psychopharmacology & pharmacopsychiatry are tags on Psychology & Neuroscience.SE.

Given this, it might not be necessary to have a separate pharmacology site on Stack Exchange. I think a way of evaluating this would be to somehow gather statistics on how often questions tagged with the aforementioned tags are answered on both Biology.SE and Psychology & Neuroscience.SE. If the statistics show that they are rarely answered, then that would show that there's practically not a Stack Exchange site to reliably have one's pharmacology questions answered.

The only statistics I have seen is the amount of times the tag has been used. But that only shows demand. It doesn't show how well or often these questions are answered. So, my question is this:

How could one gather the statistics on how often questions with certain tags are answered, and how often these questions are accepted, how often these questions are closed for being off-topic, etc.? Is this something one can ask moderators for? Is there a place where I can request such a "project" to be started? Or is this something one has to do oneself? I'm not good with coding, so I'm not sure how I'd go about having a program somehow scan the registry of questions on Biology.SE and Psychology & Neuroscience.SE and categorize all of the questions in the "answered/not answered", "accepted answer/not accepted answer", "closed/open" categories. Is this kind of gathering of statistics something Stack Exchange staff do?

1
  • 1
    I think this might not need any moderators or staff, but just a user that's very good at using the Stack Exchange Data Explorer. Since I'm not too well-versed in what is and isn't possible with that tool though, I'll leave answering this question up to one of those ;) – Tinkeringbell Apr 22 at 13:43
6

The Stack Exchange Data Explorer can help, but requires you to write some SQL. Useful when comparing a large amount of tags or sites, but the site search can do a lot too, if you use advanced search options.

Let's take the [pharmacology] tag on Biology as an example; you've probably already found out it has 384 questions.

How could one gather the statistics on how often questions with certain tags are answered

It has 70 questions without answers, and 314 questions with at least one answer.

how often these questions are accepted

It has 181 accepted answers.

how often these questions are closed for being off-topic

There are 19 closed questions; one of them is a duplicate but you can't search for individual close reasons. Note that closed questions often get deleted after a certain time; that will skew your analysis.

10
  • Note that closed questions often get deleted after a certain time; that will skew your analysis. That's quite unfortunate. Is there any way to retrieve this information? Could I perhaps ask the moderators for the total number of questions closed with this tag, and see how many were closed for the respective reasons? – A. Kvåle Apr 22 at 14:03
  • 1
    They should be able to find that out. (Whether they're willing to do that is up to them, of course.) – Glorfindel Apr 22 at 14:26
  • 2
    @A.Kvåle here is a SEDE query for some info on deletion in that tag: data.stackexchange.com/biology/query/1399977 – rene Apr 22 at 15:14
  • @rene and Glorfindel, thanks for the help! Does the tool allow for more complex operations? I realized that the statistics regarding the pharmacology tag (and the related tags) aren't really meaningful unless there's a reference. So, for example, could I make the tool do an operation like this: amount of questions unanswered with [tag] / amount of questions asked with [tag], and have it do this for all tags on the site, and get the average of all these values. That way, I'd have an average to compare to. Is this too complex for the tool? – A. Kvåle Apr 22 at 16:20
  • @A.Kvåle The tool is more or less limited only by the complexity of the query written; if you can construct the query, it should work. The only caveat is that you can get errors when trying to run expensive operation(s) on large data sets, eg. trying to do certain ops against all posts on Stack Overflow can run out of resources (Keep in mind that there are 50+ million undeleted posts on SO). But for sites other than SO, that may not even be an issue. – zcoop98 Apr 22 at 19:18
  • @zcoop98 I would only do it on the posts of Biology.SE and Psychology & Neuroscience.SE, since only the statistics within that site are relevant. – A. Kvåle Apr 22 at 19:23
  • @A.Kvåle this would answer your tag stats question: data.stackexchange.com/biology/query/1400062 notice you can download the resultset as CSV if you need further analysis. – rene Apr 22 at 19:24
  • Also worth noting that Psy & Neuro SE is called "Cognitive Sciences" in SEDE, it took me a bit to figure that out. – zcoop98 Apr 22 at 19:31
  • @rene Thank you so much! I'll mention your help when proposing the site on area51 :) – A. Kvåle Apr 22 at 19:35
  • 1
    @zcoop98 Yeah I noticed that, but I just checked what the amount of questions posted on Psy & Neuro.SE was, and the amount of questions SEDE said were posted on Cognitive Science.SE, and they both amounted to 7.2K, so I figured it was safe to say they were the same at that point. – A. Kvåle Apr 22 at 19:36

You must log in to answer this question.

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