Squaresville Welcomes You!

On behalf of the tiny, pixellated residents
of Squaresville, welcome.
We're a community of fake residents
Learning how to make a better city.

Background

This experiment was based on Agent Based Modelling course from the Santa Fe institute.

Agent-based modelling leverages non-linear or emergent behaviour of complex, internetworked phenomenon like natural processes or human social structures.

Simple models with just a few independent and dependent variables can reveal unexpected trends.

It's an interesting methodology and conducive to citizen science. The tools are:

  • NetLogo
  • R
  • Various standard operating system tools
  • dunno

How Many Politicians?

This experiment looks at how many politicians a jurisdiction needs.

It varies the number of democratic representatives, and measures income and tax burden of the residents.

For each number of representatives, the model is run several times using different random starting points.

The size of city wards varies between model runs; this is unrealistic but not a failure of the experiment.

Workflow

The experiment's workflow is as follows:

  1. Load the NetLogo model.
  2. Run a BehaviorSpace experiment.
    Emit the table results to a .csv file.
  3. Preprocess the .csv file in gawk.
    Gawk is present or available for most computers, using Brew or Cygwin.
  4. Run the graph.R script in R, importing the .csv file.
    This emits the results graphs.