Stories with longer titles get more diggs
Using the Digg API, I collected 233,570 story headlines from Digg, and the number of diggs for each story. One of the many interesting things that can be done with this data is to see how the length of a story's title is correlated with the number of diggs it is likely to recieve. The graph below shows a histogram of the average number of diggs a story recieved vs. the number of characters in its title, for all 233,570 stories, except those with more than 1000 diggs, or with a title of length 0 (I excluded these cases because the stories with very high digg counts skewed the averages considerably, and 0 length titles seem bogus).
From this graph, it is apparent that up to about 70 characters in length, stories with longer titles tend to get more diggs. Bear in mind that these numbers are averages, not medians, so stories with large numbers of diggs make the average seem unusually high. Regardless, the trend in the graph is very clear.
From this graph, it is apparent that up to about 70 characters in length, stories with longer titles tend to get more diggs. Bear in mind that these numbers are averages, not medians, so stories with large numbers of diggs make the average seem unusually high. Regardless, the trend in the graph is very clear.
posted by fbriggs at 7:39 PM
3 Comments:
Have you ever read Freakonomics? The author explores how correlations can be made between two factors, but that it doesn't necessarily mean they are dependent on each other. It's just a relation, you can't be sure if either factor affects the other without more information.
Interesting assumption here though. If the data is real, it's worth looking into.
Check out my blog if you like.
See ya round
Matt
There is some interesting stuff on Freakonomics. I used the term 'correlation' because I wanted to avoid suggesting that the length of the titles were a cause of the digg averages. I intend to post on how I gathered the data after I have had some more fun with it.
Thanks for writing this post.
Very interesting indeed!
Keep up the good work!
Best regards,
Alan.
Post a Comment
<< Home