What Tim O'Reilly, Lady Gaga, and Marissa Mayer All Have In Common | Hacker News Reader

What Tim O'Reilly, Lady Gaga, and Marissa Mayer All Have In Common

miningthesocialweb.com

8 pointsptwobrussell10 comments

With Twitter automation, I think it's an understatement to say that following count is not a reliable measure of influence. Your two subjects here have real world influence, and it's reasonable that they could arrive at those huge following #'s without gaming Twitter.

But I've noticed that certain popular tech news blogs (for example) - every single time they have a post, there are hundreds of Twitter accounts that retweet the content, instantaneously. And of those accounts, they ONLY tweet that outlet's content. I think investigating those most popular tech blogs' average "what happens when they post an article" via this type of Jaccard index would be an interesting exercise in exposing some tomfoolery, and enlightening in seeing how much stuff is being gamed. Knowing what we know about human nature, is there any doubt that people are going to great lengths to juice their follower counts and game the appearance of getting "engagement?"

The threads being pulled in this post have really interesting implications for type of analysis of "what's really going on" both in terms of a Twitter profile's following, and the authenticity of the "sharing" that occurs among their followers when they post something.

Very true. Followers is a very weak approximation. It's almost like utilizing just pagerank on a search engine. In many ways measuring influence is analogous to creating a search engine. You constantly have to search out and avoid the noise.

+1 mathattack -- in terms of understanding, I couldn't agree more about just measuring followers. It's like saying you understand a person by the volume of mail they send/receive without looking at what's in the envelope.

Really big fan of @SocialWebMining's work and I hope to see some articles about what's being said in tweets, and I think that we'd both like that a lot...

That seems a very apt analogy. The work in this post was fairly cursory since my primary goal was to "teach people how to be dangerous" as opposed to "prove" anything definitively, but once you start really working with the data, it quickly becomes so apparent that you could spend the rest of your career just in this "little" microcosm of the web.

Please add @AngelProfessor@RebelMouse to such distinguished list. Please confirm of course via LinkIn: Dr Peter Hamilton, University of Destiny, searching People not Companies.

Please consider adding Dr Peter Hamilton @LinkedIn to that list

didn't know about Jaccard index. great code examples

Some "distance-metric porn" here : http://cran.r-project.org/web/packages/proxy/vignettes/overv...

proxy::dist() in R in case you want to quickly try some of them.

Frechet distance is interesting, too:

"Imagine a dog walking along one curve and the dog's owner walking along the other curve, connected by a leash. Both walk continuously along their respective curve from the prescribed start point to the prescribed end point of the curve. Both may vary their speed, and even stop, at arbitrary positions and for arbitrarily long. However, neither can backtrack. The Fréchet distance between the two curves is the length of the shortest leash (not the shortest leash that is sufficient for all walks, but the shortest leash of all the leashes) that is sufficient for traversing both curves in this manner." (from Wikipedia)

I remember trying it out as a way of quantifying how two related stocks' price movements can diverge (just as an example application).

Nice!

Reply on news.ycombinator.com