Loading ...
Sorry, an error occurred while loading the content.

9326Re: [webanalytics] Peripheral software to support analytics

Expand Messages
  • Steve
    Feb 1 12:54 PM
    • 0 Attachment
      I've been toying around with 'R' for doing various statistical
      calculations and graphing recently. More so from a performance
      tracking perspective, but with a goal towards using it elsewhere.
      http://www.r-project.org/
      R is GPL'd and hence free. Aside to Stephane: One of your comments, I
      think on Avinash's blog, mentioned box plots? R can do those quite
      easily.

      To give an idea of how easy R is to use, here's an article I wrote
      back in October:
      http://www.stedee.id.au/Apache_WebSite_Responsiveness


      I make use of the various reports that Google allows one to pull out
      of their "Webmaster Tools". While I've not yet written a summary
      script, a goal is to grab the many months worth of accumulated data
      and see if anything useful or interesting is in there.
      We basically use these reports to identify those queries that users
      should be coming to us for, but for whatever reason, aren't. Generally
      it''s pretty obvious - slack copy.
      Google Trends is also useful as a comparator. We're getting "10
      #traffic" from word A, and 20 for phrase B, but trends shows A should
      be triple that of B! What can we do to make A better? Style of thing.


      Robbin's pointed me to the Overture keyword selector tool which I've
      found very useful. I just keep forgetting about it.
      http://inventory.overture.com/d/searchinventory/suggestion/


      You may have seen it already, but Avinash had (yet another.Ho Hum. ;-)
      ) great article on various external tools:
      http://www.kaushik.net/avinash/2007/01/five-free-advanced-web-analytics-examples-look-outside-think-different.html


      Otherwise much as Jerome, Database (in our case Oracle) and/or
      spreadsheets. The spreadsheets more for the aggregation of data over
      long periods of time.

      As I tend to roll my own tools based on need as well, I'll use the
      basic set of building blocks from a unix command line:
      [ef]?grep, cut, sort, uniq, awk or perl for more complex cases.
      There's a basic example of that in the aforementioned article in the
      section marked "pre-analysis".

      HTH?
      Cheers!

      - Steve

      Disclaimer: Despite the above repeated references to Avinash, I derive
      no financial kickbacks or other renumeration from any traffic sent to
      his web site. I am merely a Blog-Groupie of his. :-)
    • Show all 6 messages in this topic