Idea for the Group By example in the Introductory Tutorial -- The Boltzmann Wealth Model

[st346]

The Group By example in the Introductory Tutorial -- The Boltzmann Wealth Model sets up 3 ethnicities in the model, but then chooses just one at random and sets all agents to this ethnicity

    # Create a list of our different ethnicities
    ethnicities = ["Green", "Blue", "Mixed"]

    # Create agents
    MoneyAgent.create_agents(model=self, n=n, ethnicity=self.random.choice(ethnicities))

This results in a monotone chart.

On a separate issue I changed 'ethnicity' to 'allegiance', as I feel it is less contentious.

I modified my example so that agents are created with no allegiance.

    # Create agents
    MoneyAgent.create_agents(model=self, n=n)

This is reflected in the Agent Class __init__. This class also has a new choose_allegiance function defined.

def __init__(self, model, allegiance=None):
    super().__init__(model)
    self.wealth = 1
    self.allegiance = allegiance

def choose_allegiance(self):
     # Create a list of our different allegiances
    allegiances = ["Green", "Blue", "Neutral"]
    self.allegiance=self.random.choice(allegiances)

After creating the Agents the model than asks them to choose their allegiance.

    # Create agents
    MoneyAgent.create_agents(model=self, n=n)
    
    # Get agents to choose allegiance
    self.agents.do("choose_allegiance")

When plotting the histogram I added the multiple='stack' parameter to produce a stacked histogram.

  # assign histogram colors
  palette = {"Green": "green", "Blue": "blue", "Mixed": "yellow"}
  sns.histplot(data=data, x="Wealth", multiple='stack', hue="Allegiance", discrete=True, palette=palette)
  g.set(title="Wealth distribution", xlabel="Wealth", ylabel="number of agents");

My step definition of the step in the model hasn't changed, apart from using allegiance and changing 'Mixed' to 'Neutral' in the if... else...

def step(self):
    self.datacollector.collect(self)
    # groupby returns a dictionary of the different allegiances with a list of agents
    grouped_agents = model.agents.groupby("allegiance")

    for ally, similars in grouped_agents:
        if ally != "Neutral":
            similars.shuffle_do("give_money", similars)
        else:
            similars.shuffle_do("give_money", self.agents)  # This allows mixed to trade with anyone

I feel that the result is a more interesting histogram that shows the Neutrals end up poorest with 0 or 1, having given their money away, but not received any money from green or blue.

I feel this may be closer to what was originally intended in the Group by example, as the definition of the model's step can deal with a mix of ethnicities, but this doesn't actually occour in the original example.