That's absolutely correct, if this was modelling (as may well have been in other circumstances) a sexually transmitted disease, we would know perfectly well that the population does not have a "uniform pairing rate" which is a basic assumption of most of these models. However, the model will be based on the average no of transmissions per person. It is difficult but not impossible to factor in non-uniform transmission rates and time dependent transmission rates, fairly easy, but the parameters you would use would be pure guess work.
What the graphs do show is what is happening right now. And that is both important and useful for understanding what is going on and what steps we should be taking. It will also let us know when and if our measures are having an effect.