Introduction

Particle People is an interactive tool for demonstrating how an infectious disease can spread through a population. It is developed by researchers at University of Warwick and Swansea University to explore how interactive visualisations involving computational simulations can be used in educational settings. By watching the simulations, students can see how phenomena such as exponential growth and community spread arise from individual interactions. By changing model parameters, they can explore how properties of the disease or the actions that people take impact the spread.

Particle People is designed to illustrate basic concepts; it does not model COVID-19 or any other real disease. Options such as "delta" and "measles" should not be taken literally — they are used to highlight specific characteristics of diseases in an educational setting.

Simulation

image/svg+xml normal normal sick sick normal->sick recovered recovered sick->recovered dead dead sick->dead exp exposure image/svg+xml susceptible susceptible immune immune susceptible->immune (vaccine) symptomatic symptomatic susceptible->symptomatic asymptomatic asymptomatic susceptible->asymptomatic symptomatic->immune dead dead symptomatic->dead asymptomatic->immune exposure

Details

In a Particle People model, each person has a state. In the simple model, the possible states are normal, sick, recovered and dead — as shown on the model diagram. At each time step, we use the model's probabilities to update each person's state: they either remain in their current state, or transition to a new state. Most probabilites can be adjusted using the sliders, and the probabilities can be shown on the model diagram by selecting 'Show probabilities'.

Most state transitions are simple: the transition probabilities from the current state (where staying in the current state actually counts as a transition) sum to 1 and we sample a new state based on these probabilities. Transitioning from an uninfected state to an infected state is more complex. When an uninfected person is exposed to the virus (i.e. a gray 'exposure circle' touches them), we first decide if they are infected based on the relevant probability. If the person is infected, we then choose which infection state they transition to based on the relavent probabilities (assuming there is more than one infection state). Similarly, if the person is not infected, we choose which uninfected state they transition to.

Notes:

  • An uninfected person can be exposed to multiple infected people at the same time. In this case, we sample to see if the first person infects them, if they don't we sample again to see if the second person infects them, and so on.

  • Since the models only know the current states of people when choosing the next states, we can get some unrealistic behavior. For example, a person can become infected at one timestep and recover the next.

  • In both the simple and advanced models, the probability of recovering from symptomatic infection is 90%; the probability of dying is 10%.

  • A person is only exposed if the center of their particle is exposed — so the particle size does not affect the disease spread, it is just for visualisation.

  • Selecting a different model or scenario (e.g. "delta" or "measles") resets all parameters.

  • Paticle People was inspired by The Washington Post's "Simulitis" and The Guardian's measles simulation.