The Stack Overflow podcast is back! Listen to an interview with our new CEO.
2 replaced http://stats.stackexchange.com/ with https://stats.stackexchange.com/
source | link

That's a very good question, Tobias! As we're talking about StackOverFlow, I though "heck, there should be a way to see statistics from our sites"... So a googled a bit (searching for StackExchange Statistics throws you lots of results from http://stats.stackexchange.com/https://stats.stackexchange.com/, so not a good search criteria, hehe)... I found this page: https://data.stackexchange.com/

You run SQL over the StackExchange page. Just awesome!

So... digging a bit:

  • Queries with at least one answer: 2420
select count(*)
from posts
where 
AnswerCount > 0
  • Queries with an accepted answer: 1262
select count(*)
from posts
where 
AnswerCount > 0 and
AcceptedAnswerId <> 0

So, your assumption was VERY accurate. Nevertheless, I decided to play a bit with the info we could obtain here. As an Excel lover, I tried to draft some graphs to see where we can focus our efforts.

The graph below is an analysis over the total of unanswered questions x user reputation:

select p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount, count(*)
from posts p, users u
where 
AnswerCount > 0 and
AcceptedAnswerId is null and
p.owneruserid = u.id
group by p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount

Total questions unanswered

There are two clear peaks: at the near-zero reputation and near-hundred reputation.

So, now back to your original question, 'how to improve the rate of questions marked as solved', I'd say that the best we could do, as community, is to help new joiners to the community explaining the importance of marking questions as 'answered', since most of the unanswered questions (almost 80%!) are from new joiners (reputation below 50 and below 200).

Unanswered Questions Grouped

That's a very good question, Tobias! As we're talking about StackOverFlow, I though "heck, there should be a way to see statistics from our sites"... So a googled a bit (searching for StackExchange Statistics throws you lots of results from http://stats.stackexchange.com/, so not a good search criteria, hehe)... I found this page: https://data.stackexchange.com/

You run SQL over the StackExchange page. Just awesome!

So... digging a bit:

  • Queries with at least one answer: 2420
select count(*)
from posts
where 
AnswerCount > 0
  • Queries with an accepted answer: 1262
select count(*)
from posts
where 
AnswerCount > 0 and
AcceptedAnswerId <> 0

So, your assumption was VERY accurate. Nevertheless, I decided to play a bit with the info we could obtain here. As an Excel lover, I tried to draft some graphs to see where we can focus our efforts.

The graph below is an analysis over the total of unanswered questions x user reputation:

select p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount, count(*)
from posts p, users u
where 
AnswerCount > 0 and
AcceptedAnswerId is null and
p.owneruserid = u.id
group by p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount

Total questions unanswered

There are two clear peaks: at the near-zero reputation and near-hundred reputation.

So, now back to your original question, 'how to improve the rate of questions marked as solved', I'd say that the best we could do, as community, is to help new joiners to the community explaining the importance of marking questions as 'answered', since most of the unanswered questions (almost 80%!) are from new joiners (reputation below 50 and below 200).

Unanswered Questions Grouped

That's a very good question, Tobias! As we're talking about StackOverFlow, I though "heck, there should be a way to see statistics from our sites"... So a googled a bit (searching for StackExchange Statistics throws you lots of results from https://stats.stackexchange.com/, so not a good search criteria, hehe)... I found this page: https://data.stackexchange.com/

You run SQL over the StackExchange page. Just awesome!

So... digging a bit:

  • Queries with at least one answer: 2420
select count(*)
from posts
where 
AnswerCount > 0
  • Queries with an accepted answer: 1262
select count(*)
from posts
where 
AnswerCount > 0 and
AcceptedAnswerId <> 0

So, your assumption was VERY accurate. Nevertheless, I decided to play a bit with the info we could obtain here. As an Excel lover, I tried to draft some graphs to see where we can focus our efforts.

The graph below is an analysis over the total of unanswered questions x user reputation:

select p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount, count(*)
from posts p, users u
where 
AnswerCount > 0 and
AcceptedAnswerId is null and
p.owneruserid = u.id
group by p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount

Total questions unanswered

There are two clear peaks: at the near-zero reputation and near-hundred reputation.

So, now back to your original question, 'how to improve the rate of questions marked as solved', I'd say that the best we could do, as community, is to help new joiners to the community explaining the importance of marking questions as 'answered', since most of the unanswered questions (almost 80%!) are from new joiners (reputation below 50 and below 200).

Unanswered Questions Grouped

1
source | link

That's a very good question, Tobias! As we're talking about StackOverFlow, I though "heck, there should be a way to see statistics from our sites"... So a googled a bit (searching for StackExchange Statistics throws you lots of results from http://stats.stackexchange.com/, so not a good search criteria, hehe)... I found this page: https://data.stackexchange.com/

You run SQL over the StackExchange page. Just awesome!

So... digging a bit:

  • Queries with at least one answer: 2420
select count(*)
from posts
where 
AnswerCount > 0
  • Queries with an accepted answer: 1262
select count(*)
from posts
where 
AnswerCount > 0 and
AcceptedAnswerId <> 0

So, your assumption was VERY accurate. Nevertheless, I decided to play a bit with the info we could obtain here. As an Excel lover, I tried to draft some graphs to see where we can focus our efforts.

The graph below is an analysis over the total of unanswered questions x user reputation:

select p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount, count(*)
from posts p, users u
where 
AnswerCount > 0 and
AcceptedAnswerId is null and
p.owneruserid = u.id
group by p.id, ownerdisplayname, u.reputation, title, tags, viewcount, p.creationdate, answercount

Total questions unanswered

There are two clear peaks: at the near-zero reputation and near-hundred reputation.

So, now back to your original question, 'how to improve the rate of questions marked as solved', I'd say that the best we could do, as community, is to help new joiners to the community explaining the importance of marking questions as 'answered', since most of the unanswered questions (almost 80%!) are from new joiners (reputation below 50 and below 200).

Unanswered Questions Grouped