Up-vote frequency per capita ordered by country

Question

As a social experiment, I'm curious to know if users in certain countries up-vote more than others. Maybe the answers help users in some regions more than others, or maybe users in those regions are quicker to the upvote. Or maybe it's totally random.

I tried to group my upvotes by hour-of-day in SEDE, which might give an indication of trends by continent, but the time in the Votes.CreationDate is scrubbed for privacy reasons. So is Votes.UserId. Is there any way to figure this out, with or without Data Explorer?

Here's the SEDE query I started working on (and abandoned):

-- PostTypeId = 1:Questions, 2:Answers
-- VoteTypeId = 2:Upvotes, 3:Downvotes, 5:Bookmarks

declare @UserId int = ##UserId##

SELECT
  DATEPART(hour, v.CreationDate) vote_hour,
  count(1)
FROM Votes v, Posts p
WHERE v.PostId = p.Id
  AND p.OwnerUserId = @UserId
  AND p.PostTypeId = 2
  AND v.VoteTypeId = 2
GROUP BY DATEPART(hour, v.CreationDate)
ORDER BY vote_hour;

Answer 1

No, you can't figure this out. And I imagine that even Stack Exchange would need to do some serious effort in combining multiple data sources to get a rough estimate at best.

First of all: No one is required to provide a country. There is no field "Country" in the user table, there is no field "Country" anywhere on your user profile. The closest field that could be linked to / relate to country is the "Location" field.

Location is not mandatory. Only 25% of Stack Overflow users provided a Location
Location is a free text field. It has 255,862 different values.

Let's assume for a moment we're allowed to use the internal SEDE instance which has much more tables and don't have the votes table sanitized. To get the country they might be able link the vote of a user to the IP address belonging to that user. I don't have info on how much of that data is kept. Then use the IP address to do a country lookup. If we forget VPN's for a moment and assume most users are likely to use Stack Overflow from the country that match their nationality this could give you a rough guestimate. Additionally they might use Google Analytics data to enhance their data.

I think it is not possible to get an accurate vote frequency per capita ordered by country. And if we did it is unclear how that specific statistic is going to help improve voting as a mean of content quality rating. I can't deny it sometimes looks like voters didn't make the mental switch when they closed Facebook with its cat pictures and opened a Stack Overflow Regex question and continued "liking" what they see. But that is not a country specific problem that is going to be solved with statistics. That needs better upfront guidance, psychology, social behavior studies and mind reading. And probably more keyrings so you're less likely to lose your key