Exploring Your Data
Tableau can be the perfect tool to study your data by creating interactive visualizations. Below, I go through some ways to visualize your data in Tableau using the 2015 CMS Mapping Medicare Disparities data (https://data.cms.gov/mapping-medicare-disparities). First off, there are many options for extracting data from the CMS website but all of the data sets are provided separately. The good thing about Tableau is that it makes it easy to import separate data sets and either blend or join them together in tables and graphs. I look at median income and the Prevention Quality Indicator (PQI) overall composite per 100,000 population for all counties in the U.S.
Maps are very easy to create in Tableau. All you have to do is drag and drop geographical data (“County and State”) onto the Detail tab and measure values (“PQI”) onto the Color tab. Some tweaking may be necessary depending on the geographical data used. Fips codes and States will work better than county names, since there are instances of two counties having the same name (different state) you will have to tell Tableau what county/state combo is accurate. What if you have multiple data points for counties and you want to be able to compare them or visualize the difference? Here, I create a map of the PQI composite value and a scatter plot of median income and PQI composite value for every county. Looking at the map we can see that there is some regional variation in preventable admissions.
The scatter plot suggests that there is a slight inverse relationship between median income and PQI value. To take it one step further, dragging the “Urban” pill to color, I can see that there are differences in PQI between urban and rural counties. Lower income rural counties seem to experience more preventable admissions than lower income urban counties. There is also a lot more variation in PQI in the Rural counties than the Urban.
To view counties with the highest or lowest PQI value you can create a table or bar graph like the one below; however, you can only fit a small proportion of counties on the page.
Beyond standard visualization methods:
While the standard graphs and tables available in Tableau are solid and easy to use, there is room to apply many methods including calculations, parameters, etc. that go beyond standard graphs. Below, I adapted the plot as described here: https://www.perceptualedge.com/articles/visual_business_intelligence/journey_to_zvinca.pdf . It allows me to visualize all U.S. county data for median income and the PQI composite. While this may not be a production ready plot, it does provide a way to visualize both variables (~6,248 data points) on one page, along with summary statistics. Below is every county in the U.S. from lowest median income on the left to highest median income on the right. The color denotes the PQI composite score, low (light purple) to high (dark purple). You can see from this graph that higher median income counties tend to have lower admissions for preventable conditions; however, there are a few counties with high median income and a high PQI composite score. The two leftmost columns are mainly Puerto Rico where PQI composite and median income is lower than other counties in the U.S.
From here, you can modify or filter the visualization however you want. Below I filter the data by state and still have an idea of how the counties in Michigan stack up nationally. You can customize your tooltip to include the two variables in your table.
There are many other options for you to use to explore your data in Tableau, these are just a few examples I have found helpful. Let the data services team know if you have any questions.