4

So, there's been a bit of discussion as to what brings in the most viewers on this site. Trying desperately to learn about databases, I started to work on a query. With a bit help from Keen, we have this query to show for our hard work. Also, just to show how the top 20 tags are, I created another query.

So, here's a few interesting trends. The top 10 are all of the type that had one shot popular questions. Those include subjects like E.T., Middle Earth, and Pern. Game of Thrones is the first one on the list that has a fair number of questions. Also of note is that Spells seem to have a higher view count, as do suggested order.

Of the top 20 tags on the list, the most popular are: , , , , . Of particular note to me is the fact that is a very popular tag. Rounding out the bottom of the list are: , , , , and the lowest of them all, .

So, what can we learn from this? Well, the goal of this site is not to bring in large numbers of viewers, but it certainly helps. Just something to think about on getting your next question viewed by tons of people.

7
  • Those story-identification haters out there: make sure you notice the first clause in the third paragraph. Story-ID is lowest only of the top 20 tags. It's around the middle of the tags in terms of views.
    – Tony Meyer
    Commented Mar 2, 2012 at 4:33
  • @TonyMeyer - I care less about absolutes (I'm a Jedi! Yeah!), but about ratios. What is the % of total visitors that are brought in by Story ID tags? And what's the % of total visitors per Story ID tag question? On that basis, I'd expect Story ID tag to be nearly useless.
    – DVK-on-Ahch-To
    Commented Mar 2, 2012 at 13:21
  • @Pearson - I'm a bit concerned that your research does not attempt to account for external advertising. E.g. one post by Atwood/Joel on reddit/wherever is enough to severely skew the # of visitors (for example, E.T. one seems to be popular based on propaganda efforts, not necessarily its intrinsic quality/interesness, not that I'm complaining). Mind you, I'm not against the idea of popularizing the questions this way, I am merely concerned that any conclusions drawn from your excellent and great query results without adjusting for this skew might be very wrong.
    – DVK-on-Ahch-To
    Commented Mar 2, 2012 at 13:23
  • 1
    When analyzing those kind of data, I prefer to refer to the Median. I made this for you : data.stackexchange.com/science%20fiction%20and%20fantasy/query/…
    – DavRob60
    Commented Mar 2, 2012 at 13:36
  • @DVK Using the median avoid the reddited post abnormality.
    – DavRob60
    Commented Mar 2, 2012 at 13:39
  • Also of interest is the total overall, data.stackexchange.com/science%20fiction%20and%20fantasy/… . It seems that Story Identification is the #4 tag in terms of views overall, it just doesn't bring as many people per question.
    – PearsonArtPhoto Mod
    Commented Mar 2, 2012 at 13:42
  • Should we encourage users to reach out on other social media? I know I've seen questions tweeted, for example, but is there more to do? I've only tweeted one question, so I'm no champion for the cause yet. Pardon my ignorance, but it may help me AND others to strengthen the outreach.
    – Meat Trademark
    Commented Jun 11, 2016 at 20:42

3 Answers 3

Reset to default
3

Your method will not produce comparable results for franchise tags and for question type tags. Users who view one Star Wars question are likely to view other Star Wars questions, so if they might view (say) 5 questions out of 399 and thus raise the average by 0.0125. Users who view one story identification question aren't likely to go and browse other story identification question, so they'll only raise the average by 0.00266.

A better measure would be user retention: associate first views with eventual reputation, or something like that. But we can't track the first view. We could try to plot first tag against eventual reputation; even that would be a bit tricky, because it wouldn't be very meaningful for users who come through the front page and are immediately interested in a large chunk of it.

2
  • "But we can't track the first view" - are you talking as a moderator, or from standpoint of SE infrastructure? Some of SE blog posts led me to believe that sort of information might indeed be available to SE employees, though I am not certain. Great point on the most useful metric, BTW.
    – DVK-on-Ahch-To
    Commented Mar 5, 2012 at 19:03
  • @DVK We can track the point at which a user creates an account. Not the point when the person first became aware of the site. We (or rather the devs, mods have nothing resembling this) can get some idea by tracking referers to the account creation page.
    – user56
    Commented Mar 5, 2012 at 20:14
2

I honestly think trying to figure out which Tags bring the must user is pointless, specifically after analyzing the median number of views per tag query. The difference between the average and median is too big.

What type of questions bring in the most viewers?

It's good questions that will bring in the most viewers.

I separate them in two groups :

  1. Question that everybody ask themselves. In this case, viewers come from Google. The views are slowly but constantly accumulating on them. Example of one of those question are Who is Anakin Skywalker's father? or Why did Harry Potter intentionally lose the Resurrection Stone in the Forbidden Forest?
  2. Really thoughtful/insightful question or answer that got publicized. (First from the stackexchange Hot Question List then from Reddit or other) Those question got views in burst and then almost fall in oblivion after. Like How long was Bill Murray's character (Phil Davis) supposed to be in a time loop in the film "Groundhog Day"? and Are E.T. and Star Wars in the same universe?

I think the best tool to find/ analyse those questions is the greatest hits List.

4
  • Unless I missed one the first story-id question there is the poorly titles Book from the late 1980's about an abandoned space station on page 4 (around number 190), and there are a couple more on page 6.
    – dmckee --- ex-moderator kitten
    Commented Mar 5, 2012 at 10:42
  • @dmckee wrong thread?
    – DavRob60
    Commented Mar 5, 2012 at 13:05
  • Sorry. That was related to your last line there...the "greatest hits" list. There are some story-id's in there, though you have to dig a little.
    – dmckee --- ex-moderator kitten
    Commented Mar 5, 2012 at 14:42
  • @dmckee While nothing in the question indicate it's about story-id, I don't think they bring many user to the site, and data tend to indicate this. While I'm not against story-id questions, I don't think It's a good way to prove they are worthy. View count is a popularity contest; story-id is niche expertise. It's just like knowing almost all possible causes of an board error message and defining the specific causes from one's description. While it's not useful to the mass, it could useful to a restrained set of peoples.
    – DavRob60
    Commented Mar 5, 2012 at 15:45
0

As far as I understood it, Jeff had two big complaints

  1. Guessing game.
  2. Unsearchable, so their existence add little value.

These is some risk of both of these problems.

If the poster doesn't remember very much---and especially if the factoid they have is common---then we are reduced to random guessing.

And "looking for a book" type titles speak for themselves. Yuck!

But here's the thing...and we're already doing it...both problems are at least part reduced if we

  1. Require that the OP have something (preferably several somethings) concrete about the story for us to work from.
  2. Get the leading concrete bits into the title.
1
  • I don't understand the connection between the question and your answer. Did you mean to post here?
    – user56
    Commented Mar 5, 2012 at 20:17

You must log in to answer this question.

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