More Analytics

I love web analytics, and my interest was raised even further recently when I saw what Google Analytics can do.  From a development perspective it's challenging because there's quite a few obstacles to overcome: the two big ones are data storage and presentation.

For anyone who has worked on a site with any reasonable amount of volume realizes that it's almost impossible to store all data for extended periods of time.  Most web analytic software maintains this detail information for days, but before long the data is summarized for the needed reports.  The secondary challenge of presenting that data makes analytics one of the most challenging solutions to develop.

In my latest reports I wanted a way to geographically present the data.  The challenge here was twofold: one was taking logfile data and mapping it geographically, the second was plotting the data on a map.  The first question that comes to mind is, "Can you geographically map an IP address?"  And the answer is: very often, you can.  It's true that IP addresses are inherently not geographic in nature, but subnets often are. 

I like to call this the "scud approach."  It's a bit inaccurate, but a small enough map doesn't need high accuracy and the results tend to average out.  Consider that on a 720 pixel-wide map, each pixel is roughly 30 miles.  In the IPs that I manually tested, I found they were off by no more than 10 miles.  That's pretty good, but results will vary.

The image above is a thumbnail from one of the reports  -- click here to go to the live data.  In the next few blog posts, I'm going to outline how these maps are generated on the fly, and likely create some new ones along the way.  On these pages, white dots represent most users, while red dots represent the top 5 percentile -- areas with the most return visits.  (On the "All Visits" report, the red dots appear to be more than 5 percent, but it's not.  The white dots tend to just blend together and get hidden.)

This information can be a huge asset.  By understanding the geographic distribution, it's possible to map out how site content causes shift in site visiting habits.  While the other charts I've done have been valuable, these maps take the same data and display it in new ways -- and sometimes, it's just fun -- I have visitors in the Canary Islands?  Cool!
Comments are closed

My Apps

Dark Skies Astrophotography Journal Vol 1 Explore The Moon
Mars Explorer Moons of Jupiter Messier Object Explorer
Brew Finder Earthquake Explorer Venus Explorer  

My Worldmap

Month List