Age Friendly Communities
Since the beginning of 2017 I’ve been working, pro-bono, for a San Diego based non-profit whose primary focus is to cater to the civic data needs of the region. This includes supporting local non-profits, civic organizations, journalists among others in their data needs; as well as building a community of technologists that can help put data to use in ways that benefit local causes.
The San Diego Regional Data Library has been active in the region for a good part of this decade now. The Age Friendly Communities (AFC) project was setup by the data library to help CARR, Consumer Advocates for RCFE Reform, a non-profit advocating policy related to elder-care, with its data needs. RCFE (Residential Care Facilities for the Elderly) is a term used for non-medical facilities that serve individuals aged 60 and above. It is commonly referred to as “Assisted Living” facility. Learn more about RCFEs here.
Our team at the data library helped CARR collect and analyze data on assisted living facilities, current and forecasted estimates on populations with ADOD (Alzeihmer’s, Dementia and Other Diseases), demographic and income information for both the general population and for seniors, across the county. The objective was to quantify the current and future needs of assisted living facilities - ADOD among seniors being a prominent reason for assisted living use - and to assess the preparedness of communities within the San Diego county to address those needs. CARR was also interested in exploring relationships between the RCFE capacity of a community, its demographics, median household income, size of low income population among other income related information.
Below: AFC data filtered by Medium Household Income of Seniors
As someone that tackled much of the initial data exploration and analysis for the project I’ve developed a great appreciation for all the effort that goes into just finding the right data. Although demographic and income data was easy enough to find, the ADOD data was hard to come by. It was not so much brute-force searching for ADOD datasets using the right keywords on Google (although, that does work often enough) that helped; rather, it was educating oneself, even moderately, about the relevant domains (health care, senior-care) that led to the right data sources.
Below: AFC data view by Sub-Regional Area (SRA)
Even so, the latest data available (published in 2016) was for 2012. And furthermore, the data came in the form of a table embedded deep inside a PDF report on ADOD. Contrary to the hype, and despite the quintillion bytes of data we generate on a daily basis, large swaths of highly useful data in domains as critical as healthcare, transit, small business, among others are still not available to us, or are closed. It will take time and effort (policywise and otherwise) for this to change. But here in San Diego, California, and the US in general, we seem to be heading in the right direction; but never fast enough.
Below: Comparing 2030 ADOD Estimates against current RCFE capacity. Larger the bubble, higher the ADOD estimate
Given the rather diverse set of indicators we needed, the data ended up being pulled from multiple sources - a part of the RCFE data came from the CA Dept of Social Services, the other from CA Dept of Health Care Services; the ADOD data came from the County of San Diego’s Health & Human Services Agency; the demographic data came from SANDAG while the income data came from ACS.
Below: Comparing RCFE capacity by zipcode
The number of different sources used for the data presented us with challenges in terms of reconciling the most recent data available in each case e.g.: the latest ADOD data, as mentioned previously, was for 2012 while all other data was current as of 2016. This forced us to use 2012 data for the rest of the indicators to make fair comparisons of demand and supply; the windows over which specific indictors were aggregated e.g.: the existing estimates for ADOD were only available for those aged 55+, rather than the desired 65+ Note: 65 is the standard cut-off age for seniors; and the geographic units for which these data were available e.g.: the ADOD estimates were only available on an SRA (Sub-Regional Area) basis while the demographic data was available for geographies of much smaller granularity such as census tracts. In some cases we were presented with multiple options for sourcing the same data e.g.: demographic data is available through both ACS and SANDAG while income data is available through both ACS and CA Dept of Finance. There is no one-size-fits-all rule book on how to handle these issues and the best solutions require a good bit of back-and-forth with the end-users of the data and its analysis (in this case, our client CARR).
Below: Correlations
Additional Analysis
For additional analysis, the Python code used to wrangle the data, and the Tableau workbook containing the visualizations take a look at repo here.
ANALYSIS Python VISUALIZATION Tableau FORMAT csv ACCESS Direct Download