-
Notifications
You must be signed in to change notification settings - Fork 2
Description
When running a model across multiple countries and populations, the current data structure makes it somewhat difficult to access the actual data outputs directly. There are a number of different objects that are generated and attached to the Mozaic, TileSet, and/or Tile objects, sometimes in different formats (such as unlabelled matrices with simulation run outputs).
What would be especially helpful is to have a more general top-level way of viewing all model data in a top-level pandas DataFrame object. There is already an option to get this for a specific sub-model --- if I'm understanding the nomenclature correctly, this would be a single country Mozaic? E.g.:
my_df = country_mozaics_d[FORECAST_START_DATE]['DAU']['IN'].to_df()
I think what would help would be to have one default way of getting this across these objects. Maybe this means changing the default usage for generating models, or maybe it just means incorporating FORECAST_START_DATE, 'DAU', and 'IN' as columns within one overarching DataFrame.
Exactly which columns should be included in this DataFrame by default is worth discussing -- especially if there is concern that it may be overwhelming or confusing to include too many.