2
$\begingroup$

Is it possible to produce the list all mathoverflow users from a given country?

$\endgroup$
7
  • 1
    $\begingroup$ Many users do not disclose their location... $\endgroup$ Oct 5, 2017 at 20:55
  • $\begingroup$ Some probably do obscure their IP numbers. But many? $\endgroup$
    – Stefan Kohl Mod
    Oct 5, 2017 at 22:22
  • 6
    $\begingroup$ IP address history is commonly considered "personal information", so (as far as I know) it is not accessible to user-level tools. $\endgroup$
    – S. Carnahan Mod
    Oct 6, 2017 at 5:55
  • 2
    $\begingroup$ meta.SE: Finding Top Users by country. It is possible to get something from SEDE, but among the results you get only users who disclosed their location in the profile. You can look at some of the existing queries here or here. $\endgroup$ Oct 6, 2017 at 6:01
  • 2
    $\begingroup$ BTW there are also a few queries which attempt to find users from Ukraine. For example, User List: Top N - Ukraine- with false positive exclusions. $\endgroup$ Oct 6, 2017 at 6:46
  • $\begingroup$ Now I noticed a related older discussion on this site: Geographic distribution of people on MathOverflow. $\endgroup$ Jul 21, 2019 at 5:32
  • $\begingroup$ 1. Define "from a given country". How should we consider a user native from Taiwan, who works in a university in Crimea and is posting an answer from a French IP address because he happens to be there? 2. What is your goal? Why would you like to have this list (so that we can be sure this is not an XY question)? $\endgroup$ Apr 23, 2021 at 12:34

2 Answers 2

5
$\begingroup$

You have tagged your post both (feature-request) and (support)

The part can be probably answered mainly by SE staff, which could say whether there are plans to implement something like this. However, there is also a similar feature request on the main meta: Filter users by country. So there you can see how it was received and maybe eventually also some decision whether or not it is going to be implemented will be given there.

Let me try to answer the part of your question. (Which I understand as: What can be done with the existing tools without implementing new functionality?)

I think that ideal tool for tasks like this is Stack Exchange Data Explorer. (Since becoming part of SE network, MathOverflow is included in the database.) From the information about users available in the database, the Location in the Users table seems like the best choice for finding some information about country. (Of course, we will not find users which for some reason left this field blank or included something else there.)

You can create your own query or use some of already existing ones.

  • A rather simple approach is simply to search for country name in this field. Perhaps reasonable addition is adding also country code. (Country codes might lead to some false positives, but you will not miss users who list their location, for example, as "Toronto, CA".) This approach is illustrated in answers to this question: Finding Top Users by country. The query Top users by country is linked in one of the answers there.
  • Depending on how much time you are willing to spend on this, you might try to create more sophisticated queries. For example, you can try to find users who included only city in the location, but not the country. (Again, this might lead to some false positivies.) The following query, that somebody created specifically for Ukraine, seems to be quite good example of this: User List: Top N - Ukraine- with false positive exclusions. Similar query for Canada might serve as an example where excluding false positives needed more effort. (Judging by the comments at the end of the query, the creator might still consider this query to be unfinished.) It seems that the same user created similar queries for several other countries. (Since the name of queries this user created seem to be similar, another way to see them is to google for user list with false positive exclusions site:data.stackexchange.com.)

It goes without saying that you can find many already existing queries for searching by country in general or from one specific country. This is how I found the example with Ukrainian users.

Do not forget to switch sites before running the query, if you are interested in results specifically for MathOverflow.

$\endgroup$
2
  • $\begingroup$ Thank you. Indeed it works (somehow) but requires the knowledge of a programming language to create a specific data base query. Fortunately, this has been done by somebody else. I would be desirable to have such search by pressing the button "Users" in MO and selecting the country somewhere. $\endgroup$ Oct 10, 2017 at 5:41
  • $\begingroup$ If you want to see something like that implemented, upvoting the feature request on meta.SE is a reasonable thing to do. (Although personally I do not see this as something very likely.) As I mentioned, my post was an attempt to address the support-part of your question (rather than the feature request). $\endgroup$ Oct 10, 2017 at 15:23
0
$\begingroup$

Below is the result of this SEDE query. It's significantly more complicated than the one martin mentioned in his comment, largely because it was only composed within the last year and Martin's comment was from 2017 (though the link in his comment does have this new SEDE query listed in one of the comments too).

enter image description here

Like all the others have been saying, a SEDE query can only be as good as the information it has available, and in this case, for example, 3 users seem completely missing from the SEDE result:

enter image description here

Those three users do have their locations given in their profile, so it might be possible to scrape this data somehow and further improve the results of the above SEDE query, but we will always be limited by flaws and gaps in the information available.

$\endgroup$
6
  • $\begingroup$ As far as I can tell, you have used a query from one of the answers here: How can I write a query DATA.SE to list top users from all countries in the world? In the same question, rene's answer shows how that query can be modified. For example, you can tell the query that Lviv should be counted as Ukraine or Oxford as UK. (Of course, this will no longer be correct if the top user is from Oxford, Maryland.) $\endgroup$ Apr 23, 2021 at 5:48
  • 1
    $\begingroup$ Here is a query with such minor modifications: data.stackexchange.com/mathoverflow/query/1400175 $\endgroup$ Apr 23, 2021 at 5:53
  • 1
    $\begingroup$ As the example of Noah Snyder indicates, the query seems to mistake the state of Indiana (in the US) for the nation of India. $\endgroup$
    – S. Carnahan Mod
    Apr 29, 2021 at 2:00
  • $\begingroup$ @S.Carnahan Interesting, the query seems it could benefit from some refinements for sure. For the most part I think it's fairly useful though. $\endgroup$ Apr 29, 2021 at 2:02
  • $\begingroup$ @S.Carnahan That's exactly the kind of problem that the query in my comment tries to address. (Although there is no doubt that there are other improvements. And sometimes the Location fields is not enough to tell - for example University of Georgia can be in the USA or in Tbilisi.) You can check what you get after those modification for India or ... $\endgroup$ May 16, 2021 at 12:50
  • $\begingroup$ ... USA or some other country. Since I have mentioned Georgie I will add also such link. $\endgroup$ May 16, 2021 at 12:51

You must log in to answer this question.

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