Getting a Good Job Depends More on Race and Gender than Education by Urban Institute

This data tool published by the Urban Institute features unique radial visualizations categorizing 108 occupations in terms of quality and occupational crowding across demographic groups. It begins with a detailed scrollytelling walkthrough explaining the research and data. As the user moves down, they can select an occupation and see how the job's quality score was built, then scroll through a series of cards showing how each demographic group is crowded into or out of that particular occupation. Each card is lightly annotated to remind the user how to interpret the radial visualization.

This project was built in SvelteKit using LayerCake and d3-force. Individual occupations are organized into concentric rings in the radial visualization to reinforce the crowding theme of the data. For job-quality data, Urban researchers identified 11 key indicators from publicly accessible data to represent job quality related to occupational crowding and combined these indicators into one dataset. To understand how these variables differed by occupation, researchers looked at occupations at the four-digit Standard Occupational Classification (SOC) code level.

To compare job quality among occupations, researchers used the values for all above variables to create a job-quality score. If an occupation's value for a variable was better than average for all occupations aggregated, the occupation received one point. If the occupation's value was worse than average, the occupation received zero points. A lower value indicated a “better” score for two indicators: unemployment and injury rates. Researchers then calculated a final score, which is the total number of elements for which an occupation had a better-than-average value. By this process, researchers interpreted a higher total score as a better overall job.