annatz / opinion-dynamics-simulation Goto Github PK

This project was developed for the 'Data Programming in Python' course, part of the MSc in Data Analytics at the University of Glasgow in 2023. It presents a comprehensive simulation framework designed to explore the dynamics of opinion formation within communities. It models the interactions between individuals with varying stances on a given issue, leveraging spatial positioning and social networks to simulate opinion shifts over time. The simulation is implemented in Python, structured around a robust object-oriented design that prioritizes computational efficiency, especially critical for scenarios involving a significant number of individuals and/or time steps.

• Class Definitions:

  • Individual: Base class for all individual types. This class encapsulates the core functionalities and attributes common to all individuals in the simulation, including:
    • Location: The spatial coordinates (x, y) representing the individual's position.
    • Opinion History: A record of the individual's opinion changes over time.
    • Contacts: A list of other individuals that this person is connected to within the simulation.
    • Update Method: A method to update the individual's state based on the simulation's dynamics.
  • UnbiasedIndividual, NegativeIndividual, PositiveIndividual, StubbornIndividual: These classes inherit from Individual, each implementing specific behaviors and initial conditions reflecting their predisposition towards the issue at hand.
  • Simulation: The central class that orchestrates the simulation, incorporating methods to initialize the population, run the simulation, and generate outputs for analysis.

• Key Features of the Simulation Class:

  • init: Initializes the simulation with a set of parameters and creates the initial population.
  • run: Executes the simulation over a specified number of time steps.
  • plot_network: Visualizes the state of the population at a given time point, showing individuals' opinions and their connections.
  • chart: Generates a time series plot showing the evolution of opinion groups within the population.
  • most_polarised: Identifies the maximum polarization observed in the simulation.
  • most_polarised_day: Determines the day on which maximum polarization was observed.
  • individual_summary: Produces a detailed breakdown of each individual's opinion changes over time.
  • friendship_similarity: Assesses the opinion similarity among connected individuals.
  • friendship_similarity_chart: Charts the evolution of opinion similarity over time.
  • ensemble_statistics: Runs multiple simulation instances to analyze variability and statistical properties of the model outcomes.

• Simulation Parameters:

  • T: Number of simulated time steps.
  • N: Number of individuals in the simulation.
  • αpos, αneg, αstub: Probabilities that an individual is of the positive, negative, or stubborn type, respectively.
  • γC, γop, γdist: Parameters influencing the likelihood of connections based on spatial proximity and opinion similarity.
  • βupdate, βspread: Control the rate at which individuals' opinions change and the influence of differing opinions.
  • γextr: The probability that an individual becomes completely convinced (opinion shifts to 0 or 1).

Ensure Python 3.x is installed on your system. The required Python packages are listed in requirements.txt.

1. Clone the repository:

   git clone https://github.com/AnnaTz/opinion-dynamics-simulation.git

2. Navigate to the cloned directory and install the required packages:

   pip install -r requirements.txt

Launch Jupyter Notebook or JupyterLab and open opinion_dynamics.ipynb. Execute the cells sequentially to explore the simulation setup, run scenarios, and analyze outcomes.