My Photo

« Something For The Weekend | Main | If You Like This Blog, You Might Like These »

December 15, 2006

Comments

Dibyo

Just wanted to say that I enjoy your blog.... (it also feeds into my google reader).

Also, I thought I'd mention that wouldn't you count feeds as a stronger measure of influence than just reading? i.e. if you had to make a metric of how 'strong' or 'popular' a blog is, you need to count the number of people who feed it, and the number of people who read it, but give a heavier weight to the feeders than the readers since feeds are a measure of loyalty.

Did I get my point across?

Moshe Koppel

Matt,
That actually looks to my naked eye like a pretty decent correlation. What's the R2 of a linear regression?

Moshe

Sean Tierney

Maybe I'm missing something but this scatterplot seems to be the expected result to me... it shows correlation between subscribers and inlinks, just not a strong one. But think about the people you read regularly, do you link to them proportionate to how often you read them? Probably not. The people I follow regularly are consistently on point with their insights but I tend to write more about the stuff I run across from my own daily experiences in working and observing or from the "needle in the haystack" I find on del.icio.us. This is a great idea for an experiment but this is about what I would expect. Can you elaborate as to what you were expecting to see?

sean

Matthew Hurst

Moshe - R2 is 0.4861, not that good.

Rama

That looks like a great correlation to me. R-sq of 0.48 is pretty good considering the nature of data you are trying to measure. And especially if you want to compare influences rather than absolute measures, you have found a cheap proxy. It does not cost much to measure inlinks and that as a surrogate of the influence looks valid, IMHO.

Good luck on that and I would love to see more analysis on that one.

Rama

Matthew Hurst

Put it this way, in the range of 100-110 inlinks, feed subscription ranges from 5 to 13, 939. I just pulled this range randomly (and I'm sure the 5 is anomolous in some way). Perhaps one way to measure this would be to count the number of different positions if one used inlinks as a rank to using feed subscribers. One could also measure the obvious distribution of deltas in between the two rankings.

The comments to this entry are closed.

Twitter Updates

    follow me on Twitter

    March 2016

    Sun Mon Tue Wed Thu Fri Sat
        1 2 3 4 5
    6 7 8 9 10 11 12
    13 14 15 16 17 18 19
    20 21 22 23 24 25 26
    27 28 29 30 31    

    Categories

    Blog powered by Typepad