I have data that is of the ordinal flavor where some fields have rows with numbers 1 - 5 representing user satisfaction levels as well as the text N/A for not applicable.
Is it better to treat these columns (fields) as text or number data types, and likewise for columns that have nominal numbers 1 and 0 representing yes or no? My instinct says they’re ordinal/nominal therefore text.
Which brings me to the next question of whether it is useful (in Power BI) to check for correlations between these types of ordinal variables (fields) or not? If it is useful, then perhaps N/A could be changed to a numeric value (maybe 0 or -1)?
I am considering using the Correlation Plot visual in Power BI. It requires numeric values. I welcome all suggestions but would like to also know the best practice.