Is it a must to use synthetic key in my case?

[kwhwan]

I'm developing an app using 3 different tables, which consisted of the task status from 3 different teams under the same department. Each day I will receive excel tables from each team and I will add timestamps (businessdate, something like a snapshot date) to each table to record the status of each tasks.

As a result, the data model is accumulative, i.e. the old records will not be deleted so that I can view the status of each task as of each snapshot date. A simple illustration of the data will be like this.

Table 1:

Table 2:

Table 3 is similar from another team.

The objective of this app is to create a consolidated report combining the data from the 3 teams.

I will need a filter for "businessdate" so that I can view the status of each task as of different dates.

Another filter will be a but more complicated - "Completed YearMonth", I will join the 3 different completed dates (i.e Task Completed Date and Order Completed Date in my examples) and create a YearMonth dimension for them, so that I can check all the orders and tasks done within the same month in my department

That means, there will be synthetic keys between the 3 tables as there are multiple fields joining:

1. businessdate 

2. Completed YearMonth

I have done a lot of research but still cannot find the perfect solution for my case.

I have tried to create a "canonical date" for the dates but it is creating a circular reference

If the joins of the 2 date fields are intentional as I want to use 2 filters only for the 3 tables, may I know if it a must to use synthetic key in my case?