Data interaction: You can zoom any region of the network with your mouse and click on any node to visualise the neighbourhood for that character.

The social graph of Breaking Bad is created (code here) by scraping the description of each episode as published on Breaking Bad Wiki. The relations are based on the textual copresence of characters in any scene. The scene synopses were text mined for instances of character names, that is, if a character is only named in a scene but not physically part of the scene the script will fail to understand the difference.

I manly use the graph to test community detection algorithms but then the data (you can download it in JSON format here) can be applied to all sort of analysis of social graphs. And of course you can spend few nice minutes exploring the ego-network of Hector “Tio” Salamanca…

The visualisation above is rendered by Sigma (a JavaScript library for graph visualisation) with a script downloaded here and adapted by me.

Hopefully further developments will include some dynamic modelling of the network. I want to answer the following question: Is there any chance to predict who will die during each season based on relative network positions?

Tweet about this on TwitterShare on LinkedInShare on Google+Share on FacebookEmail this to someone