As we can see, endorsements are indeed closely correlated with total downloads. While there is some variation, there is not a lot. It could be more interesting to see how the percentage of people who endorsed the mod compares to total downloads. We do not have a field that provides this percentage, so we will have to create a new field. We can do so by dividing Endorsements (number of people who have endorsed the mod) by Unique Downloads (the number of people who downloaded the mod). This will work because a user can only endorse a mod once, and they can only do so after they have downloaded the mod. Unlike total downloads, unique downloads will not be affected by any subsequent downloads by the same user.
Learning Objectives
- Create a calculated field.
- Duplicate a worksheet.
- Format axis number display.
[Endorsements]
will be shown.
Technically, typing "Endorsements" with brackets is not particularly complicated, but I would not have remembered the required format without dragging and dropping Endorsements. Most likely I would have wanted to default to how Tableau references fields in the tooltip. If you are confused as to why the required format for the tooltip is different from that of the calculated field, consider what is being referenced in each case. The tooltip is not referencing the field from the dataset itself, but rather the field as it relates to the particular worksheet. For example, the tooltip might provide the sum or average endorsements, calculations that will be defined by aspects of the worksheet. Is the sum of endorsements the sum of all endorsements or the sum of all endorsements for a given category or country? A calculated field, however, is not unique to the worksheet we are on when we create it, which is why we did not start by creating a new worksheet. It must therefore reference the field itself.
[Endorsements]/[Unique Downloads]
.
Now that we have a field to plot, we need to create a new scatter plot. As there is no need to reinvent the wheel, so to speak, we will duplicate our current scatter plot instead of starting from a blank slate. It would be a shame to have to redo our formatting for the new plot.
Our approval ratio scatter plot is still showing endorsements and total downloads. We want to replace endorsements with our new calculated field. One method of doing this could be to remove Endorsements from Rows and then add Approval Ratio. Alternatively, we could add Approval Ratio first and then remove Endorsements. We will go with an even simpler option.
The plot will have a very different distribution than the Endorsements scatter plot. Instead of diagonal line indicating strong correlation between total downloads and endorsements, we see that approval ratio has quite a bit of variation with low download values, then levels off.
We have a few things that we need to clean up after replacing Endorsements with Approval Ratio as some of our formatting options have been undone. Fortunately, most of the formatting we applied in the first scatter plot remains, so we do not have too much work to finish this one.