For the past month I have been working on a Twitter graph project. I took inspiration from 's "Mapping the Furry Fandom" (youtube.com/watch?v=IDzIko), and implemented it to be real-time, and work with any user.
Link: graph.bunnypa.ws
Conversation
The algorithm: Generates a matrix between all your friends, and counts how many people each friend-pair are followed by in common.
Then it finds each user's top friends, and connects them (up to TopN).
3
6
43
The list on the left is the count of common followers.
App does not work on phones.
Please be kind to my server. This is just a pet project. And give any feedback. :D
Source code can be found here:
3
24
Very cool congrats. What would it take to bump up the limits past 10,000 nodes? Could you reach 80K? This would allow a lot more people to give it a try.
1
The problem is three fold.
Calculating the proximity matrix of all those users would take many server hours.
And then having to load like a GB file json file into the browser.
And then the browser would have a hard time handling that many objects.
It is possible, but hard...
very nice! didn't change many options and it still had a decent shape after a short time
1
1
Show replies
Show replies
Interesting! I hope my paleo-following butt doesn't cause too much consternation.
1