1 of 1 people found this helpful
I believe your problem would be solved by blending in a file with the text response (e.g., "Strongly agree") along with the numeric response. Steve Wexler talks about this in his blog: http://www.datarevelations.com/reshaping-survey-data-with-tableau-9-0.html. He also recommends using a helper file to group questions, which will make it easier to select the questions you want to visualize together. That post starts with info on reshaping; if you're already familiar with that, skip down to the text vs. numeric and helper file explanations.
Hope that helps!
This looks like it might be exactly what I need. I'll take some time to review it and hopefully get this method to work with my data. Thank you!
What survey platform are you using? Does it happen to be Qualtrics? If so, they now have a data connector to Tableau that makes all analysis WAY easier. If not, then I also recommend using Steve Wexler's info on how to format your data for Tableau. The only downfall to this is you have to do this each time you want to update the data. If you have an ongoing survey, I am not sure it would be helpful. His website www.datarevelations.com is really helpful.
Yes, the data is collected with Qualtrics! However, it needs to undergo a very extensive data cleaning process (including appending additional data and running syntax to assign additional demographic fields) in SPSS. We are considering the data connector for our other, less complicated survey projects, however.
Have you worked with these specific instructions before? When I try this method, the two columns of data (labels and numbers) do not align after reshaping. It looks like the Tableau reshaping add-in is excluding null cells within the labels, but not excluding them from the numbers. Without recoding all null cells in the data labels as “null,” do you know of a way to correct or work around this?
I am sorry I don't. The data sets I have all worked with didn't have null values. If there was no answer for a particular question it just left it blank instead of needing it to say null.
I have not had the issue you are describing with the Excel add-in tool, but null values in my data sets are typically represented with a period, rather than just a blank cell. I have occasionally noticed unexpected results when connecting to the data in Tableau if there are null values in the first few rows of data. My solution to this has been to just sort my data before connecting in Tableau such that there are no null values in the first row. Perhaps this would work for you with the Excel data transformation?
I had more or less the same problem. What I recommend is to have this kind of database processing in SPSS or some other statistical software first. You could easily create text variables from numerical ones. Once you have your text variables you can pivot the data and all that, being much easier to work with it. Of course this can take you time, but if you're using SPSS the sintax can help you through this repetitive task of replicating variables.
My point is that if you're familiar with other statistical software (SPSS, SAS...) you could take advantage of this by getting your database ready and clean before moving to Tableau, where I find this kind of task much more difficult to perform.
I had similar problem when dealing with SPSS .sav file. I decided to go manually using Calculated field:
Case [Question Name]
WHEN 1 THEN "Strong Agree"
WHEN 2 THEN "Agree"
Above calculation assumed that [Question Name] is measure, if it is string then use "1".