Research talk:Anatomy of English Wikipedia Did You Know traffic

Interesting, on a quick initial read. It's a pity you didn't tag for art, which usually has a reasonable number, & I think tends to score rather better than average. Also for music - though I suppose pop etc are in "popular culture" - are they? Or does that mean all "culture"? Not clear. I usually check the views of my (now 215+) entries & find that, beyond whether or not it has the image, the innate interest of the subject & hook make the biggest difference - I'm not suggesting you can analyse these subjective qualities. Biographies considered by date would be interesting. How do contemporary people compare with say WW2 and then older periods. More on overall DYK traffic would be interesting, if that has not been done elewhere. What are the average total views per day or per set, and how do they vary? Johnbod (talk) 13:20, 3 August 2013 (UTC)

I think it was more of an issue that the contributors I looked at did not write art related DYKs, so they were under represented in the sample. The problem is the data was manually compiled by hand, and it just took forever to get. (And if you look at the actualy DYK inclusion list, I did not always include the time of day, in a number of cases because that added a took a lot of time to manually get.) My raw data is there and easily available for re-use so if you have a list of DYKs about art, another analysis could be done. Culture for me was different that popular culture. Culture included literature and non-popular culture music and not so much television, pop music and movies. I would have to look at the raw data again to see what was included under that tag exactly. But yeah, there is a lot of volume for research in this area to better understand patterns. It would be nice to see some one take the initial data set or develop a more comprehensive set on their own to examine things. MilHist type articles were tremendously over represented by Gibraltar related DYKs and pretty much did not reflect much diversity than that outside some UK things. Lots and lots of variables. --LauraHale (talk) 17:07, 3 August 2013 (UTC)

My own personal observations confirm those of Johnbod in regard to hook placement entirely. The Did you know statistics (DYK STATS) list shows, that the majority of new entries with over 5,000 views, in any given month, are illustrated. This paper would benefit greatly from the analysis of those results; more specifically, by looking at what lead hooks (i.e. illustrated with thumbnail pictures on the front page) actually failed to reach the minimum number of hits set up as benchmark for the DYKSTATS placement, with probable cause. Poeticbent talk 14:54, 3 August 2013 (UTC)

Would be nice to see that. There are a number in the sample that failed to reach the 5,000 views point. After spending about 10 days playing with the data and trying to write up the results, it led to more questions than answers regarding the causes for traffic patterns related to DYK. It would be a really interesting area to see more investigation into. --LauraHale (talk) 17:07, 3 August 2013 (UTC)

Thanks for your prompt response, LauraHale. For the sake of impartiality, what percentage of your leaders (mentioned) did not make it to DYKSTATS? Poeticbent talk 00:22, 4 August 2013 (UTC)

The fourth note is erroneous. It states If an administrator has not moved a holding area to the queue, a bot will automatically pull the most recent holding area to the queue. This is viewed as not desirable for a number of reasons. This is not true. Prep areas must be moved to the queue by human administrators. All that the bot can do is move queues to the main page. --Orlady (talk) 00:44, 4 August 2013 (UTC)

Ooops. Removed footnote. --LauraHale (talk) 07:53, 4 August 2013 (UTC)

German DYK Analysis

Hello, I did a similar analysis on German DYK articles and their traffic in January 2012. The data set looked like this. My summary of the results appeared in the German signpost. I carried on the analysis for two more years, as you can see from the backlinks to the data set. --Minderbinder (talk) 20:57, 16 August 2013 (UTC)

Very cool. :) --LauraHale (talk) 17:26, 19 August 2013 (UTC)