Data Modeling Question

[Anonymous]

Hello, I need advice with a data modeling problem. Below you may find 3 screenshots and attached PBIX file.

Link to Download PBIX file

Screenshots below show:

1. What my report looks like
2. The problem I am experiencing
3. My current data model

This data model contains several linking tables for many-to-many relationship between Recommendations table and different geography levels tables (State, County, City). I know that some relationships need to be changed in order to fix this problem but I'm not sure which ones.

Screenshot 1 - What my report looks like:

p

Screenshot 2 - The problem I am experiencing:

p

Screenshot 3 - My current data model:

p

Thank you!

[Anonymous · Accepted solution]

I was able to figure out this problem on my own. In order to take care of hierarchy between geographical dimension tables (State/County/City) all I had to do was to create Location table with nullable Foreign Keys. See below screenshot.

[amitchandak]

@Anonymous , I removed some bi-directional relations. But this model needs changes. recommandation will merge with State_rec , country_rec and City_rec and create three fact recommandation_State_rec  , recommandation_country_rec

City_rec_recommandatio. They will not join with each other 

https://www.sqlbi.com/articles/the-importance-of-star-schemas-in-power-bi/

[Anonymous]

Thanks for looking into this. 

I have tried your proposed solution but it breaks cross filtering across other visuals. See below screenshot.

Can you please elaborate on how you propose to change the data model. I am afraid I don't fully understand your previous explanation. I need to keep cross-filter both directions for State > County > City relationships.

Also I assumed state_recs, county_recs and city_recs are already considered fact tables since they contain foreign keys from recommendations and states/counties/cities respectively.