December 31, 2012

Duck! Here comes another year!

Our most-viewed posts in the past year, according to the Jetpack backend

  1. Tip of the icecube: on the ‘hundreds of unfit teachers’
  2. Who is really buying New Zealand: dramatically bad bubble graphs
  3. A story from last year, about the suicide rate at the Foxconn plantĀ 
  4. David Scott’s Super 15 predictions for round 2
  5. Another one from last year, on lottery odds.

This only counts hits directly on the page (eg from Twitter or RSS feeds), not from browsing the site, so it misses about 1/3 of the traffic.

The lottery post shows an interesting pattern of hits

Big Wednesday

The second burst of popularity was a building jackpot, with the spikes on Tuesday/Wednesday each week. There’s a similar pattern on Google Trends for the phrase ‘Big Wednesday’.

 

There were 31 posts in the new ‘Denominator’ category, which mostly follows violations of two of the most basic numeracy principles for reporting:

  1. Auckland is bigger than the other cities, so simply having more of something in Auckland is not news
  2. If you have two years of data, either take an average or a difference. Don’t report a total.

 

Approximately 95% of our comments were spam, but the filters caught 99.8% of it. The 5% of real comments are much appreciated.

And finally, the title of this post is from an Ogden Nash poem.

avatar

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »

Comments

  • avatar
    Nick Iversen

    I’m curious about the spam and surprised that it’s as high as 95%. Is this spam human or bot generated? Would a CAPTCHA reduce it?

    There could be some fun CAPTCHAs implemented for the site. A simple question such as “correlation does not imply …” would weed out all bots and half the mechanical turks (humans that are paid to solve CAPTCHAs) and most journalists (sorry, couldn’t resist that jibe).

    11 years ago

    • avatar
      Thomas Lumley

      Since the filters already catch nearly all the spam, it’s not really worth having CAPTCHAs. We already moderate all posts by new commenters, and that’s more reliable than a CAPTCHA. The only reason to switch would be if the hand-approval got too slow.

      11 years ago