Typically, survey data is formatted so that each row corresponds to an individual respondent and a column for each question. This results in what we like to call ‘short and fat’ data, where we only have as many rows as respondents, but many columns for all the questions. However, Tableau prefers that data be ‘tall and thin’. To adhere to this preference, we need to pivot the data so that we’ll have fewer columns and many more rows.
When you pivot your data, you want to keep any dimensions—that is, the fields you want to ‘slice and dice’ by—out of the pivot so that they are stored as separate columns independent of ‘Questions’. With survey data, this tends to be any demographic information about your respondents. For example, age, gender, country, etc.
This results in your many columns being converted into just two ‘Pivot Field Names’ (renamed to Questions) and ‘Pivot Field Values’ (renamed to Responses). This multiplies the number of rows that were originally in your data source by the number of questions included in the pivot. Now, each row in the data set corresponds to one question per respondent.
There will often be ‘null’ responses, meaning a respondent didn’t answer the question. You should apply a data source filter to exclude these so that every record corresponds to an answered question per respondent.
You can learn more about pivoting your data in Tableau with this Quick Start Guide.