What is the best strategy to create a basic model?

Guys, another simple question. I would like to know the best strategy to create the model below.
You can consider that we have an excel file with one tab for each table.
The main questions are below:
Should we create a dataset for each table and then make an Analyzes relating to all datasets?
After creating the Dashboard, how do I maintain the Dashboard? Can I only do this by publishing another Dashboard?
If I delete the Analyzes (after publishing a Dashboard), will the Dashboard still work as it is connected to the Dataset?

Thanks guys.

Hi @woliveiras, you can join the tables in the data prep phase and create a new dataset combining all tables into one, to use in the analysis . QS currently don’t support multiple tables with relationships added to analysis. You can add multiple datasets in the analysis, but they are treated as individual tables.

For you second question, yes you are able to delete the analysis and still have the dashboard linking with the original dataset. If you accidently delete the analysis, and want to get it back, you can click save in the dashboard and save as a new analysis.


Amazing explanation @emilyzhu . Just to make sure that I understood. To create a data prep phase, I need others tools (ETL tools), right? Considering QuickSight package, we do not have tools to help in the data prep phase. Am I right?
Thank you so much.

1 Like


this will help you

1 Like

Hi @Naveed , thanks for sharing. This is exactly what I would do. the question here is:
In that case, Am I loading the same file three times? (I know that we are using different tabs, however I am concern about load sequence.
Another question: is it possible to create a degenerate dimension based on the flat table?


You can generate the calculation field in dataset as we doing on analysis.
It’s is good to done all the logic on your source/raw data.

Naveed Ali