Better to Pivot or UnPivot?

[ELW]

In general, is it better to have data pivoted or unpivoted if everything else is equal?  

I have an existing dashboard based off of survey data.  The old data is structured so that each response is a row, with each survey question as a column.  This works fine for most visuals.  

As we're starting to use Snowflake, the survey data has been loaded and I am working on connecting the existing dashboard to the same data in Snowflake.  It's set up in a transposed / unpivoted format where questions and responses are in rows.  I'm not sure if this is a Snowflake thing, or a decision that was made.  

My question is what is the best way to transform and load the data into PowerBI for loading speed and efficiency.  I believe I could either:

• Pivot the data into columns for each question, like the old data (lots of columns, less rows)
• Leave it unpivoted, and update each visual to filter results for the relevant question. (few columns, more rows)
• Split it into multiple tables for sets of questions. 

[Greg_Deckler]

@ELW 99.99% of the time, you want unpivoted data coming into Power BI.