An interesting approach on how to solve data modeling problem.

Istead of going from source data to target data mart, you write your models backwards, starting from the desired result and then performing actions and steps to achieve it.

1. Prototype with GSheets
2. Mock out the final report you’re trying to build
3. Mock the query that would produce this result
4. Iterate
5. Match this up with source data
6. Move the logic into your pipeline orchestration tool


Full post from Claire:

https://discourse.getdbt.com/t/writing-models-backwards-an-unorthodox-approach-to-data-transformation/2287