Creating a Dataset from Qualtrics Data

Best practices for consuming Qualtrics surveys

SAP Analytics Cloud can access data that has been provided through responses to a Qualtrics survey. When you're creating a model or dataset, it will be helpful if you have access to that survey in Qualtrics, so that you understand the structure of the survey.

Here are some suggestions to help you get the best out of analytics on Qualtrics surveys.

Note that questions in the Thrash section of your survey shouldn't be used when creating your model or dataset.

During the data preparation step while creating a model or dataset, the following basic initial steps will help in getting better insights:

1. Add a count column in SAP Analytics Cloud.

   Create a calculated column with the value 1 in each row, to count your responses.

2. The first row appears with descriptions of questions. This row should be removed.

   Why is this row provided at all?

   If you don't have access to the Qualtrics survey while creating the model or dataset, since the column headers provide the question numbers, the description information will help you decide which questions are relevant for analysis.

3. Create a Geo dimension.

   Use the columns LocationLatitude and LocationLongitude to create your Geo dimension. Video: Enrich Geographic Dimensions

Modeling different types of questions

Qualtrics offers several options for question types that may be included in a survey. Here are suggestions for handling some question types in SAP Analytics Cloud.

Net Promoter Score (NPS)

For a question of type NPS, for example Q1, Qualtrics generates a field Q1_NPS_GROUP with the value Promoter or 3, Detractor or 1, or Passive or 2 for each response. To calculate an aggregated NPS in SAP Analytics Cloud, calculate the counts of Promoters and Detractors, and calculate the NPS as: (Promotercount/count - Detractorcount/count)*100.

Matrix Table

For a question of type Matrix Table, columns are generated based on the number of statements. For example, if Q2 has 3 statements, then the generated columns are Q2_1, Q2_2, and Q2_3, and the descriptions are available as mentioned in basic step 2 above. You can create a model or dataset for each matrix question, by selecting the columns ResponseID and Q2_1 to Q2_3, doing an unpivot on the Q2 columns, and then adding a count as mentioned in basic step 1 above.

Link this dataset to the other dataset for this survey through ResponseID.

A similar approach can be followed for the Rank Order question type.

Text Entry

For these questions, if there is a TextIQ license, then performing text analysis generates some sentiment related fields. It is useful to have calculated measures based on the Sentiment score field, which can have Positive, Negative, or Neutral values.

Highlight

Avoid importing this question type during model creation because it generates a large number of columns, which could exceed the maximum number of columns supported by SAP Analytics Cloud. For these kinds of questions, use datasets instead.

Adding descriptions for dimension names and dimension members

The question numbers are available as dimension names. To have dimensions look more meaningful, add a description for each dimension in the model or dataset, based on your Qualtrics survey. For dimension members, the code values are provided. You'll need to provide descriptions based on the Qualtrics survey. You can get all of this information using the “Export Survey to Word” option in Qualtrics.

Limitations

• CSV export size limit = 1.8 GB: Currently, response exports that exceed 1.8 GB will fail. To prevent your export from failing, use limits and filters to limit the size of your final export file.
• Each API token can run a maximum of 15 jobs per minute.