Inspired by a Tableau Workout Wednesday from last year, we are going to make a Marimekko chart. A Marimekko chart, also known as a Mekko chart or mosaic plot, is used to visualize categorical data across two variables. Both axes use a percentage scale that determines the width or height of a bar. Think of a Marrimeko chart as a combination of a 100% stacked column chart and a 100% stacked bar chart. These charts are often used to provide a picture of an industry or customer base by segment.
This is a lesser used chart type because it’s really only appropriate to provide an overview. It’s difficult to compare segments when they do no share a common baseline.
Our Marrimeko chart is visualizing the percent of women in different job levels. The height of each segment of a column is determined by the percent of that job level that correspond to the selected gender. The width of the column represents the percent of jobs that are categorized as the selected job level. Since we have grouped gender into women and other, and placed the women category (which is the focus) at the bottom, this chart doesn’t suffer from some of the design and analysis limitations mentioned above.
You can complete this workout using the Microsoft-created custom visual, Deneb, or any other custom visual that supports creation of mosaic plots.
- Retrieve the data from data.world.
- Transform the job level values so they have only the first letter capitalized when they are not an initialism (VP, SVP).
- Calculate the percent of men in each level in a Power Query custom column.
- Unpivot the data so that Gender is a column and all percent values are in a single column.
- Create a mosaic plot that contains Job level on the x-axis, and gender as the series.
- Ensure that each job level is labeled clearly. (Hint: if you are using the Microsoft custom visual, you may need to use another visual to make labels.)
- Create a visual that explains in text what percentage of jobs are classified as that level and what percentage of that job level are comprised of the selected gender, based upon the selection in the mosaic plot. If no segment in the mosaic plot is selected, the visual should instruct to user to select a segment for more information.
This week’s data is stored in data.world:
You can use the data.world connector in Power BI (recommended) or download the Excel file to your local machine. You will need to log in to data.world to retrieve the data, but accounts are free.