Sunday, October 05, 2008

CS: Impact statistics - part deux.

This entry is an update to a previous post on readership statistics as we crossed the 6,000 visits in September. I have listed the previous month's stats next to the August numbers.

Currently, in order to provide information to this blog and attendant pages:
  • there are about 500 pages being crawled every day,
  • I have on average between one and two good discussions with one of you or a researcher outside of this community on the subject of compressive sensing per week.
  • There have been about 306 (prev. 286) entries on Compressive Sensing, most of them featuring extremely recent preprints.
On the readership's side:
  • ~34 (prev. ~30) people are reading this blog directly in their e-mail box,
  • 112 (prev. 102) people are reading this blog through Google Reader, while another 75 (prev. 60) maybe reading this RSS feed thanks to Feedburner.
  • ~ 218 (prev. ~180) visitors/day or about ~6500 (prev. ~5000) visitors/months (not unique) are coming to the site with about half of that traffic from the search engines and from wikipedia. The other half is people who are reading the blog through an RSS reader and who want to look at past entries.
  • The Compressive Sensing LinkedIn group has 53 (prev. 24) members since its inception a month and a week ago.
  • The readership and linkage to Nuit Blanche have enabled it to reach a PageRank of 5, while the Big Picture site has a PageRank of 3.
  • Some people come back often to either the blog or to the Big Picture site and some people stay for long periods of time (see the stats below gathered over a year and a half time period)

On the impact side, I have had some widely differing accounts. This mostly stems from the fact that:
  • I link directly to papers and Google Analytics or some other counters do not do a good job at counting those hits,
  • Sometimes Nuit Blanche is not the only blog/site pointing to the page or paper, hence it is difficult to find out the source of the increase in traffic.
  • People read the blog through e-mails, rss readers, Google cache or through the site itself, I can see only certain types of click-through to the papers.
For all these reasons, I currently estimate that about 60 to 100 (same as previous as I don't have a better way of doing this statistics) people will click through to the paper listed on the blog the day it is posted. I also estimate that over time that number is doubled. But the most important thing to remember is that once it is featured either on the Rice Compressive Sensing repository page or on this blog, Google knows about it and it becomes visible much before it eventually is published on a journals' sites.

No comments: