In this week's challenge, we focus on visualizing survey data, specifically, data gathered in the form of a Likert scale. Likert scales, named after Rensis Likert, involve closed questions with a sliding scale response format.
In this exercise, we will employ a diverging bar chart to depict the distribution of responses gathered from the survey. The chart will provide an insightful visualization of the survey results.
Dataset: The challenge utilizes a standardized sample survey dataset accessible on GitHub. This dataset serves as a common foundation for participants to work with, ensuring consistency and comparability across solutions.
Process: The following process is implemented in order to build a dashboard in Power BI to fulfill the challenge:
1. Data Transformation in Power Query:
- Promote headers
- Unpivot queries
- Create a duplicate table and assign sentiment to each query as below:
- Remove duplicates by right-clicking
- Assign an index to the data
- Generate a list of negative sentiment items
- Create a custom column as per requirements
- Merge tables to create a new master table, including Sentiment information
- Assign numerical values (1-5) to the query answers using a conditional column, considering the sentiment
2. Create a Power BI Desktop table with the following measures:
- Positive, Neutral, and Negative Counts. (Note: Neutral values are to be split into two to diverge across the y-axis, and Negative Count will have a negative value)
- Create a Total value that applies to each query using the ALLEXCEPT filter to enable % calculations for each query group
- Calculate % Scores for Positive, Neutral, and Negative Values using the DIVIDE function
- Formulate an NPS Score as (100 * [Positive Score] + [Negative Score])
- Create a Rank Calculation to order the bar chart and NPS by the NPS score
3. Visualize in Power BI: Click Here for Live Interactive Dashboard
