Computational Modelling of Primary School Segregation (COMPASS)

This Python implementation tries to model school choice and resulting school segregation based on the work of Schelling (1971) and Stoica & Flache (2014).

NB: This repo is folked from and improved based on the original GitLab repo, where the dataset used can be found here.

Usage

Install and Run

In a terminal, navigate to the project folder, install all the dependencies by running

pip install -r requirements

and run the program within a GUI by running

python run.py

This will start a local server to automatically open your web browser and show the program as a webpage. If not seeing the webpage open, manually input http://localhost:5004/ in the link bar.

Update Documentation

Install pdoc3 if you haven't already done so. Browse to the compassproject folder in your terminal and run pdoc3 --html --force --output-dir docs compass. The documentation should be updated now.

Overview

The repository consists of:

• run.py: a script that runs the model interactively with a visualisation (browser)
• testrun.py: a test script (work in progress)
• household.py: the household class
• student.py: the student class
• neighbourhood.py: the neighbourhood class
• school.py the school class
• allocator.py: allocates the students to their school of choice
• agents_base.py: overarching agent used for inheritance
• model.py: initialises the entire system and all of its components
• parameters.py: contains all the parameter values for the simulation
• scheduler.py: takes care of the activation, sequence and placement of all agents
• visualisation.py: browser based visualisation
• utils.py: containing all measurements
• functions.py: containing some math functions to be used by the classes

Simulations

Information on how to run the code here.

Testing and development

Setup a virtualenv with the required dependencies.

$ python -m venv env
$ . env/bin/activate
$ pip install -r requirements.txt

Install the package locally (in developement, or editing mode):

$ pip install -e .

Then run the tests with:

$ pytest

Profiling

Some profiling result can be found in this notebook. Also, some scaling graphs can be found in this notebook.