I'm really into learning Julia and Agents.jl (mastering these is my top summer goal), but I have really come around to loving NetLogo -- I wanted a stochastic block model with visualization, and made this from scratch in literally 5 minutes.
@edhagen Nice. I’ve been really impressed with Agents.jl, but it’s always faster to not have all that overhead.
@psmaldino In case you haven't seen it, check out this Julia package for organizing your simulations:
https://juliadynamics.github.io/DrWatson.jl/dev/
@psmaldino I used to assign reviewing simulation platforms as part of the lit review for dissertations, so came to netlogo that way. 90% chose it, remainder built the whole simulation from scratch in their favourite language (c, python, & yes once scratch). Though it’s been almost a decade now since I’ve had cs dissertation students.
@j2bryson in terms of pedagogy for social modeling, NetLogo has no real competitors, especially if you want embodied agents or structured populations. I generally encourage the grad students in my lab to use another tool for their research projects (which usually involve ABM). Most Python (with or without Mesa) or Julia (Agents.jl). I first learned with Java and MASON, which I don’t teach or encourage anymore.
@psmaldino I did an ABM in Julia using Agents.jl, and also rolling my own. My own, which was just modifying rows in a matrix, was about 10x faster than Agents (but without all that you get from Agents, of course). For mine, I think I could update 1e6 rows/sec to 1e7 rows/sec. Anyway, I was impressed.