Posted October 9, 2009on:
UPDATE 9/10: Insight into how this works and the rationale behind this in this post.
So after my last post I received a couple of graph requests.
Here’s @map_maker‘s, see how I’ve added the directionality. This one took a long time to build. She talks to a lot of people who talk to a lot of people…
This is my boyfriends, @theAlMan. He doesn’t talk to that many people, so I dared to go to a depth of 2.
My very popular neighbor, @snookca (depth 1 again). He caused a stack overflow…
@Circuitbomb (depth 1).
@uOttawaWISE – we’ve not been on Twitter long, but our network is growing! Also – because this one is fairly small you can see the connections really nicely.
@cheth (another one to cause a stack overflow!):
@sparkyourart (another stack overflow!):
Sorry for the delay – I’ve been tweaking my algorithm to produce more balanced graphs. It now takes your last 200 tweets and last 100 mentions (due to search limitations this means there will not be any older than a week) and finds the oldest ID in each set. Then it takes the maximum of these and ignores any tweets older than this ID. I’ve also added node distance coloring, so that the central node is darkest and nodes further from the central node are lighter.
Here’s how my graph looks now (let me know what you think!):
So, @velvetescape and @velvetconnect come next. After hitting the API limit a couple of times (even after I’d tweaked the algorithm – @velvetescape talks to and about a lot of people) I had to do these a little differently. For @velvetescape I changed the depth to 0 so it’s just him and the people he talks to and who talk about him. For someone with this much conversation going on, though, I think it shows it pretty nicely.
For @velvetconnect I just changed the number of recent tweets I request from the API. Normally I get 200 most recent tweets and the last 100 (or week) of mentions but this just wasn’t feasible so I cut these both in half.
@jdemond‘s is below. This one is nice – you can really see he’s part of two distinct networks, and that the network to the right has some sub-networks going on.
@KristinaThorpe – this one was interesting! I didn’t hit the API limit but I did run into problems with the visualization engine that prevented it from rendering – not sure why, but I think it may be due to the graph being very densely connected. I reduced the API data by 25% (last 150 tweets, last 75 mentions) and that fixed it!
If you want your graph done, let me know via Twitter or in the comments and I’ll put it up here for you. Feel free to use the image anywhere you’d like (ask me for a higher quality one if you need to), but please link back to this! Thanks 🙂