An introduction to agent-based models: simulating segregation with Python

If you have also used NetLogo, how different/difficult was it as compared to that.

I had used NetLogo a few years back in college to study tax evasion behaviour from agent based models, it worked out quite well to match IRS figures with similar policy variables with not much time spent in coding up the logic. But a major problem I faced then was dealing with a few matrix operations, for which I had to write out a lot of code. So I figure it would be fun to replicate it in python now, hence the question to gauge difficulty in porting that code to python...