Advanced November 2014 Statistics


So that little bit I added in my last statistics post, tracking how many days went between the first and the last reading of an article according to WordPress’s figures? I was curious, and went through my posts from mid-October through mid-November to see how long the readership lifespan of an average post was. I figured stuff after mid-November may not have quite had long enough for people to gradually be done with it.

I’d expected the typical post to have what’s called a Poisson distribution, in number of page views per day, with a major peak in the first couple days after it’s published and then, maybe, a long stretch of exceedingly minor popularity. I think that’s what’s happening, although the problem of small numbers means it’s a pretty spotty pattern. Also confounding things is that a post can sometimes get a flurry of publicity long after its main lifespan has passed. So I decided to count both how long each post had between its first and last-viewed days, and also the “first span”, how many days it was until the first day without page views, to use as proxy for separating out late revivals.

Post Days Read First Span
How To Numerically Integrate Like A Mathematician 45 8
Reading the Comics, October 14, 2014: Not Talking About Fourier Transforms Edition 25 7
How Richard Feynman Got From The Square Root of 2 to e 41 4
Reading The Comics, October 20, 2014: No Images This Edition 5 5
Calculus without limits 5: log and exp 25 3
Reading the Comics, October 25, 2014: No Images Again Edition 28 2
How To Hear Drums 14 6
My Math Blog Statistics, October 2014 30 4
Reading The Comics, November 4, 2014: Will Pictures Ever Reappear Edition 9 6
Echoing “Fourier Echoes Euler” 12 5
Some Stuff About Edmond Halley 11 2
Reading The Comics, November 9, 2014: Finally, A Picture Edition 11 4
About An Inscribed Circle 13 5
Reading The Comics, November 14, 2014: Rectangular States Edition 15 1
Radius of the inscribed circle of a right angled triangle 12 5

For what it’s worth, the mean lifespan of a post is 19.7 days, with standard deviation of 12.0 days. The mean lifespan of the first flush of popularity is 4.5 days, with a standard deviation of 1.9 days.

I suspect the thing that brings out these late rushes of popularity are things like the monthly roundup posts, which send people back to articles whose lifespans had expired weeks before; or when there’s a running thread as in the circle-inscribed-in-a-triangle theme that encourages people to go back again and again. And I’m curious how long articles would last without this sort of threading between them.

Author: Joseph Nebus

I was born 198 years to the day after Johnny Appleseed. The differences between us do not end there. He/him.

4 thoughts on “Advanced November 2014 Statistics”

  1. I noticed there are two different kinds of posts (of mine): Those that ‘decay’ as you describe – and about 5-10 that slowly get more and more popular over time.
    Or I see the combination: An initial decay in the first months, then after 6 months the posts gets more and more hits. This can be due to a backlink from a popular site (my most popular post is my review of that QFT book as the author linked back from his site) or because I obviously have included the “right” search terms. Due to whatever reason people search for “sniffing router” or the like and my post on network sniffing is gradually becoming my top post – months after the initial decay.
    And then there is seasonal popularity: Every year in October or November my post about mice getting in microwaves through the rear-side vent has a peak. In summer people search for tomatoes and espalier. So I conclude most of my readers are from the Northern hemisphere :-)

    Like

    1. I really only have the one perennial now, my guide to all the different kinds of trapezoids that I could think of. (I’m happy nobody has pointed out an obvious other one that I failed to list, so far.) Now and then one of the Price Is Right posts makes a comeback, and I suspect that the pair about what grade you need on the final to pass the class should be seasonally successful.

      I have spotted a couple of people searching for my Arthur Christmas posts too. I should make a page that collects all those into a reasonably seamless whole for better fundability.

      Liked by 1 person

  2. hmmm… i read a whole bunch of articles by merely scrolling down your home page. wordpress would note the fact that i reached your home page and this particular one but it wont tell you what else i happened to linger on ;)

    in any case, articles that address a very peculiar technical issue seem to get really high number of hits and articles that talk about sex, religion, politics and drugs are also top sellers. unadulterated mathematics or physics, as i see, does not interest the masses to a very great extent. but the bunch who do seem to care about such topics are rather dedicated ^_^

    Like

    1. This is true. Properly speaking the most popular post I ever have is my home page, but since that’s always there it seems trivial to include. And there’s no way for me to guess what people are lingering over. I’d imagine it to be the most recent couple of posts, but that’s just a simplifying assumption.

      I suspect that my best chance for really well-read posts is to come up with a bunch of slightly mathematical questions — “What is an ansatz?” “What are conjugate variables?” “What is a symplectic integrator?” — and go through each of those. I don’t suppose it’s a coincidence my most popular post is “How many trapezoids are there?”.

      And, alas, my subject matter doesn’t really get me much excuse to talk about politics or religion or drugs or sex that get so thrilling. (Well, they can, but they end up being the religious controversies of mid-17th century England, which rouse the passions relatively slowly these days.)

      Like

Please Write Something Good

This site uses Akismet to reduce spam. Learn how your comment data is processed.