Modelling sales for two cities

Questions v5
,
9

From what I understand of the problem so far, you also want to:

  • add in your two observed time series of advertising spends as MutableData nodes
  • run that through some linear or non-linear function to model the effect of advertising on sales
  • add those advertising-caused sales to the mu’s

Along with @lucianopaz’s suggestion, that should pretty much solve your problem.

You can then explore counterfactual situations of what-if advertising spend had been zero (or something else) using the kinds of methods in this example notebook Counterfactual inference: calculating excess deaths due to COVID-19 — PyMC example gallery