long tail of friends

friendpix-long-tail.png

Another nerd-alert. (Hm, two in a row, this isn’t boding well for general consumption –ed. (hey waitaminute, who is this –ed. guy and how did he start commenting in-line on my blog? –alex) ah, the joys of being the author and editor –ed.)

Ok, that little tourette’s episode aside, what you are looking at is a graph representing people in the pictures I’ve taken from 2000 through 2007. There are about 3100 pictures, and each bar represents the number of times someone appears in a photo. As you can see, it forms a long tail distribution quite nicely, although it doesn’t die out quite that fast.

One thing to note is that the top three occurrences have been artificially clipped to limit the count to 200, to make the graph easier to read. Had I not done that, the graph would have truly resembled a long tail distribution, as the top three people really skew the picture.

Enjoy.

3 Comments

  1. Jeff — October 10, 2007 #

    How did you count them?

  2. alex — October 11, 2007 #

    Lightroom counted them all for me.

    Unfortunately, I didn’t have a clever way to get that data out of LR and into Excel, so I just typed it in by hand. Kinda dumb, but it got the job done.

    I think LR uses sqlite as a backend, but I didn’t feel like grovelling around and possibly corrupting my db.

  3. Jeff — October 11, 2007 #

    Are you telling me that you are such a metadata semantic tagging fervent disciple that you typed in names for 3100 pictures?

Leave a comment

viagra