A meta-population model looks at population mobility on a large scale. It divides the study population into groups according to age group, place of residence, etc. Each group follows an underlying mathematical disease-diffusion model in which individuals are assigned to subgroups (compartments) of susceptible, exposed, infected or cured individuals who are (temporarily) immune to the virus, just as in the [compartmental transmission model](https://covid-project2022.vercel.app/article/how-can-we-model-viral-transmissions-compartmental-model).

By mimicking interaction between these different groups according to a range of scenarios--for example, (not) a lot of contact with people outside the family, or mobility between communities--the model is able to estimate via short- and long-term projections how the number of infected people will change.

Based on census data, we estimate all daily commuting movements adults make between Belgian municipalities. Those movement patterns map contacts at home (in the evening) and at school or work (during the day) by municipality.

Just like the compartmental transmission model, the meta-population model is described by a series of stochastic calculations, one for each 'patch' or geographical unit in the model, in this case, per municipality. Complementing the basic SEIR model, in this case a term expressing who enters or leaves municipality x from which compartment (susceptible, exposed, infected, cured) is added each time.

The strength of this type of model depends largely on reliable information about the number and type of social contacts that people in different age groups make and people's movement behavior between different regions, and on how measures would affect that behavior.