Speed optimization question

Can you show us some screenshots ?

The standard approach would be to have one or more "fact" tables that contain dates and amounts and multiple "dimension" tables that link to one or both of the fact tables that contain information about the "facts" that you use to filter on.

You should try for a "star" schema, where you do not have joins between multiple dimension tables. Often the "lookup fields" in dimension tables are denormalized and just hold the actual value, rather than a foreign key. Sometimes fact tables are denormalized, the table of order lines has the information from the order table added to it, for example.
I would think that the engine would be fastest working along relationships... looking things up in code would probably be the most expensive way.

I don't know if you have dax studio...it can show you execution times for your queries and is a big help in optomizing.