Hello everyone,
I have a dataset that tracks what occupation a person holds after graduating with a degree in psychology. The current sheet I'm trying to fix is a word cloud that shows the occupation titles by their frequency.
Variables:
- Occpm
- Type: String
- Details: Broader classification of occupation. (e.g. "Science & Engineering Related Occupation")
- Estimated Count:
- Type: Float.
- Details: The survey weight for a row of data. Summing these values for all instances of a specific occupation should estimate the total number of graduates working within that occupation.
- Rounded Count:
- Type: Integer.
- Details: A rounded version of Estimated Count (e.g. 400 instead of 416.33)
- Year:
- Type: Integer.
- Details: Year associated with data (e.g. "2023")
- Level of Highest Degree in Psychology:
- Type: String.
- Details: "Bachelor's Degree", "Master's Degree", or "Doctorate/Professional Degree"
- Occupation:
- Type: String.
- Details: Name of occupation title. (e.g. "Psychologists")
To create the word cloud, I switched the Marks to Text, moved Occupation to Text, and moved Estimated Count to Size. I used Occpm to Color the occupation names. Then I added Rounded Count, Year, and Level of Highest Degree in Psychology to the Tooltip. I also created a Table Calculation on Estimated Count to get % of Sum.
My problem is that the resulting word cloud doesn't seem to be sizing the occupation titles by their relative size. See the screenshots below. "Other health occupations", which represents 1% of people working in 2023 after graduating with a doctoral degree in psychology, is the largest text on screen, even though "Psychologists - Health Services" represents 39% of the total sample. What I want is for the size of the occupation text to be proportional to its percent of the total weights for that occupation (% of SUM(Estimated Count))
I've tried creating a new measure that's just a Table Calculation for Percent of All, and I tried using that for the Size pill, but that doesn't give me what I'm looking for, either.
What's curious is that, when I check on the table calculations for % of Total, the visualization temporarily looks exactly the way I want it to. But when I exit the table calculation menu, it snaps back to its original state.
I'm guessing there's either a filter or sort at play that's mucking up the works. That, or I'm missing something very simple. I can post the Workbook, if requested.
Thank you for your time and advice,
Hi Cory,
The problem is the bunch of text is large - e.g. its making the available space larger, but then the font size is adjusting to fill the available space. Optically this is giving the result that Other Health Occupations is huge, but this is just because its using less words than the Psychologists.
One technique (and a quick hack) would be to just reduce the text of Psychologists blah blah to just 'Psychologists'.
A second technique would be to use the SIZE() function - say perhaps rank your occupations and then allocate the size of the font used based off this value. This could give extreme results though, and you would need to adjust around this (e.g grouping different ranks and giving them a font size and not an individual calculation) - the complexity increases significantly for only a slightly similar outcome.
A third technique would be to encase your text in boxes or have a slight border so that viewers can see the boundary.
HTH
Peter