These are the 397 members of James Alliban's "tech art" Twitter list (plus 11 handles my algo thought should be on the list), mapped into an influence network.
Obvs, the concept of "influence" here is solely within the context of Twitter.
30,654 connections have been plotted. Each is scored according to who follows who, for how long, how often they chat, who is liking what, who is retweeted, who is ignored, etc...
Importantly, this algo scores evidence of influence felt, rather than influence exerted. This method accentuates the influence of friends and colleagues, and plays down celebrity.
The process can be applied to any Twitter list. If you know of an interesting scene you'd like mapped, make a list and ping it over to me.
For more on the thinking behind my methodology, read on...
Influence is a subtle beast. It is often barely felt, let alone measurable. But we know it is there, because we can see its effects at scale. Careers are built on this stuff. It is the dark matter of social media.
In Marketing circles, many "influence" scores are effectively "reach" scores. Simply because reach is much easier to measure.
"Reach" is the number of eyeballs a message can, yerknow, reach. Basically, it is the volume of the voice - number of followers, number of interactions, etc... This is the celebrity model, which will always rank some actor or pop star as more important than a close friend or colleague.
Celebrity models make sense for reach. But they miss a lot when it comes to influence.
Imagine Kanye were to show up to your community knitting circle. He would doubtless command a lot of attention. And any contribution he made to the discourse would be listened to with great interest. He would reach everyone in the room very effectively. But this doesn't mean he instantly becomes a knitting "influencer". Kanye knows shit about knitting.
The knitting influencer in that room is more likely to be Mrs Wooley, who has wielded a pair of needles for 70 years and knows how to use them. She will be listened to on the subject of knitting much more than a confused rapper who's clearly wandered into the wrong room. On a normal day she too would have a high reach in that context. But measured in relation to Kanye, her voice is tiny.
This is one of the problems with measuring social media influence. Social is dominated by loud voices, so the influence signal is often drowned out by the noise of reach. Influencers are usually seen as those who speak the loudest, rather than those whose words carry the most weight.
Genuine influence, the stuff that makes us change our behaviours, purchase products or vote for leaders, is more likely to be exerted by experts, or trusted friends, than celebrities. In the smaller interactions; friends recommending stuff to friends; experts reporting back on their field. The most influential tweet is not necessarily the one that gets the most RTs.
This makes it very hard to measure effectively. The subtle beast requires a more subtle approach.
My approach is based on one simple principle. Instead of attempting to measure influence exerted, I instead measure influence felt. Listening, not speaking. This is something that's possible to do with a closed network.
Traditional celebrity model methods have difficulty with this, because context-specific networks are complex and heavy to compute. Especially if Kanye, and his 30 million followers, have to be factored in.
But, by shifting the focus from speaker to listener, the number of Kanye's followers is no longer relevant. We only care whether Mrs Wooley, or the rest of the circle, are paying attention to him.
There are plenty of visible indicators of this. If Kanye were influential upon knitters (he might be, for all I know), they would follow him, like his tweets, share his videos, etc... Just as they all follow, like and RT Mrs Wooley for her rants on the Portuguese Flip vs the Right-hand Throw.
The size of Kanye's reach is evidence that Kanye is influential on the subject of Kanye. But we'd only know he is influential on the subject of knitting if he is disproportionately listened to by knitters.
This is a measure in which Mrs Wooley will score high, but Kanye may not.
This is how the connections in "Who Are Your Influences?" are calculated. A script gathers public data on every node in the network, cross references interactions to build the edges, then ranks the measure of influence back onto the nodes. The network itself is then chucked away, along with all the Twitter data gathered, leaving just the nodes displayed to the left here.