Stamen’s collaboration with UCSF takes on new urgency in the age of COVID-19.
Kelly Morrison wrote the first draft of this post
Nine months ago, a team from UCSF School of Medicine Dean’s Office of Population Health and Health Equity led by Dr. Debby Oh, hired us to build an online data visualization tool to help researchers explore how our environment influences our health at a local level across California. The goal of the project was to give researchers new insights into the social determinants of health. The data was vast, even for us: powered by numerous publicly available data sets, and curated by UCSF into compelling data stories. On our side, Logan Williams and Alec Burch built and designed the tool and Vinay Dixit managed the effort. The project is live at http://healthatlas.ucsf.edu/ as of early April and is being updated daily.
As the project was nearing completion, COVID-19 began its relentless march across the globe. In this terrifying moment, the team saw an opportunity: to help us understand COVID’s spread and health impacts through our communities by leveraging the existing dataset on the Health Atlas.
The interface was designed with health researchers, policy makers, and community advocates in mind to allow exploration of over a hundred social determinants of health and health metrics across the state at the census tract and county levels. Let’s say you’re interested in understanding age and food insecurity in the Tenderloin (where I live with my family). The tool lets you examine that neighborhood through different lenses: food insecurity (screenshot below; 24.9% of people living in the TL are food insecure), population age (23.4% of people living in the TL are under 18), percentage of three- to four-year-olds in preschool (48.7%, compared with 69.94% for the whole City), and a huge list of other demographic metrics like access to an automobile, some college or more, and so on.
The Health Atlas allows us to learn a great deal about this neighborhood and California with just a few simple interactions. Above, for example, I can tell from the bar chart that the Tenderloin (highlighted in orange) has significantly higher than average CA food insecurity (24.9% as opposed to the statewide average of 12.7%); that the Tenderloin is one of a relatively few neighborhoods where food insecurity is that high (since the bar where it lands is relatively small compared to the average); and that there are municipalities in California where food insecurity is a shocking 32.0% (and in some outlying cases even higher).
I can also quickly see on the map that the areas with the most food insecurity in San Francisco are the Tenderloin, SOMA, Chinatown, and the Western Addition (since those neighborhoods are darker on the map), and that most of the rest of northeast side of town is much less food insecure (since those neighborhoods are lighter on the map).
All of this gets wrapped up in a data story that the UCSF researchers put together about the relative landscapes of two neighborhoods in San Francisco, the Bayview and Tenderloin. I’d heard of “food deserts” before, where people have to travel long distances to get their groceries; the Bayview is one of these. New to me is the term “food swamp,” where there’s lots of food close by, but much of it is less healthy; the Tenderloin is one of these.
The Health Atlas also allows you to add a second variable for comparison. Here we’ve got the number of single adult households in blue and extremely low income households in yellow for the Tenderloin. This gives us the ability to easily see places where those two variables are both very high, since blue + yellow = green. By this measure, the TL (and some surrounding areas) pop out of the map in vivid contrast to the bluer areas around them. And the scatterplot in the lower-right corner of the screen, which treats every census tract as a dot mapped against the two axes, shows the highlighted area as the average of all census tracts in that grouping.
Here’s the scatterplot in an expanded view:
The COVID-19 pandemic has continued to reveal that illness affects our communities in California unequally. In addition to the disparities of who is able to shelter-in-place and who isn’t, it’s clear even from the preliminary statistics that rates of infection and death affect people differently due to a multitude of socioeconomic and environmental factors. As more local data becomes available and as the tool continues to update daily, we hope that the Health Atlas can help us better understand the inequities underlying the COVID-19 pandemic and work towards a future where all Californians can be healthy.