I'm fairly new to Tableau. One of the areas that I'm exploring is the analysis of internal survey data. A couple of months ago, when I was first trying out Tableau, I had an XLS containing survey response data, with one row per user - it had many columns, which could be grouped into 2 categories: demographics and survey response data. I separated the data into what I thought made logical sense, 2 "tables" (Excel worksheets), with a User ID field that tied them together: a Demographics Table and a Survey Response Table.
Bringing these 2 tables into Tableau, I obviously had to create 2 separate data sources. Things seemed to work mostly ok as I built out some sheets and dashboards but I ran into some issues, which I don't recall the specifics of right now. I know that I had tracked down a few of the issues to the fact that each sheet seemed to have a "primary" data source, and there were some limitations or restrictions based on the primary data source that is chosen, with regards to non-primary data sources included on a sheet.
Flash forward to now, and I'm starting "fresh", although the survey data is similar. I want to ensure that I put the data in the very best shape in Tableau, so I'm hoping to get some input from the group here.
I am familiar with Tableau's Excel Data Shaper add-in, and I can see how it can help, however I'm wondering if there are any best practices for shaping & pulling XLS survey data into Tableau when you have a number of columns of demographic data and a number of columns of survey response data.
I've seen examples, where it seems that the demographic data isn't shaped, only the survey response data is shaped. With a number of demographic data columns, it seems like the demographic data is being replicated unnecessarily. Besides making that overall data set larger than necessary, I've seen this approach cause some challenges, with the need to add a "distinct" clause to every chart/visualization.
So, how do you all do it?
Separate data sources for demographics and survey response data or a single data source?
Shape only the response data or shape the demographics data too?
Any other tips or tricks?
Thanks in advance!