[Time for a bit of housekeeping. For Thing 10, Task 2, I posted a tweet referring to an assignment in a completely different MOOC. Below is the content of the "explanatory Doc" referred to in a subsequent tweet, in case it becomes uncoupled.]
========================================================================
========================================================================
I figured it was unfair to tweet a “teaser,” so here’s the follow up. While pursuing Rudaí 23, I’m also enrolled in “Dazzling Data Visualization,” a free MOOC offered by the U.S. National Network of Libraries of Medicine, Southeastern/Atlantic Regional Office. For that class, I had to create a visualization based on a public data set. Living in the boondocks and being a liaison to some clinical departments on campus, I was drawn to the opportunity to visualize a large data set about federally designated “Medically Underserved Areas.” What follows is partly edited from the context I provided for my classmates, who get to critique my visualization.
-------------------------------------------------------------------------------------
I wanted a simple data set that would be accessible when visualized, so it was necessary to reduce the sprawling original set to less than 1,000 data points. (Tableau Public counts “marks” in legends as well as in the data itself.) Most of the measures used to determine Medically Underserved Area (MUA) status were incompletely populated, so I abandoned those criteria as a focus for the visualization. After much fumbling, I realized Tableau could handle minor civil divisions as “cities.” And so I decided to follow the guideline (invented by me) of “Visualize what you know,” and focused on the three rural states of northern New England.
Hoping that consistency would make for better comparisons, I used filters to get:
Designation Type of Medically Underserved Area and Medically Underserved Area - Governor's Exception
Designated status with no Break in Designation
Medically Underserved Area/Population (MUA/P) Component Geographic Type Description of Minor Civil Division
Common State Name of Maine, New Hampshire, and Vermont.
This retrieved a manageable set of 221 records.
I discovered that Minor Civil Division Name lacked some values, and turned to Medically Underserved Area/Population (MUA/P) Component Geographic Name. This still made for a lot of work. First, my current home state of Maine has a lot of minor civil divisions needing special handling in Tableau. For Maine's "unorganized territories" (and one Native American reservation) I even had to resort to using the GeoNames server to find coordinates, and then a USGS converter to transform them into decimal format. Second, there are some names in common between states. I suspect this is what accounts for some of my missing values, but a) I couldn’t figure out how to fix the problem and b) 4 missing values do not change what I eventually discovered.
With all the missing data, I wasn’t even sure what to focus upon in the visualization. So I started with something easy: the 2 different types of MUAs. Bingo! The raw data is not sorted by state, concealing what turned out to be a striking pattern: for some (unknown) reason, most of Maine’s MUAs earn their status “by the numbers.” Adjacent New Hampshire, on the other hand, requests MUA designation in the majority of cases.
To make the map as accessible as possible, I used both color and shape to identify designation status. The colors are Tableau's suggestion for color blindness. (They check out on the Coblis color blindness simulator.)
In theory, there are ways of overlapping legends to create the illusion of 1 legend capturing color and shape. In theory, there are ways of using IF statements to conceal tooltip display of a variable with a null value (e.g., Infant Mortality Rate, for many of the MUAs on the map). I tried to follow the directions on assorted blogs and guides and Tableau help screens, I failed, and after several hours, I had to cut my losses and present something decidedly less than ideal. That’s my visualization story, and I’m sticking to it!
-------------------------------------------------------------------------------------
The final product my classmates will critique is here if you’re curious.
-------------------------------------------------------------------------------------
I wanted a simple data set that would be accessible when visualized, so it was necessary to reduce the sprawling original set to less than 1,000 data points. (Tableau Public counts “marks” in legends as well as in the data itself.) Most of the measures used to determine Medically Underserved Area (MUA) status were incompletely populated, so I abandoned those criteria as a focus for the visualization. After much fumbling, I realized Tableau could handle minor civil divisions as “cities.” And so I decided to follow the guideline (invented by me) of “Visualize what you know,” and focused on the three rural states of northern New England.
Hoping that consistency would make for better comparisons, I used filters to get:
This retrieved a manageable set of 221 records.
I discovered that Minor Civil Division Name lacked some values, and turned to Medically Underserved Area/Population (MUA/P) Component Geographic Name. This still made for a lot of work. First, my current home state of Maine has a lot of minor civil divisions needing special handling in Tableau. For Maine's "unorganized territories" (and one Native American reservation) I even had to resort to using the GeoNames server to find coordinates, and then a USGS converter to transform them into decimal format. Second, there are some names in common between states. I suspect this is what accounts for some of my missing values, but a) I couldn’t figure out how to fix the problem and b) 4 missing values do not change what I eventually discovered.
With all the missing data, I wasn’t even sure what to focus upon in the visualization. So I started with something easy: the 2 different types of MUAs. Bingo! The raw data is not sorted by state, concealing what turned out to be a striking pattern: for some (unknown) reason, most of Maine’s MUAs earn their status “by the numbers.” Adjacent New Hampshire, on the other hand, requests MUA designation in the majority of cases.
To make the map as accessible as possible, I used both color and shape to identify designation status. The colors are Tableau's suggestion for color blindness. (They check out on the Coblis color blindness simulator.)
In theory, there are ways of overlapping legends to create the illusion of 1 legend capturing color and shape. In theory, there are ways of using IF statements to conceal tooltip display of a variable with a null value (e.g., Infant Mortality Rate, for many of the MUAs on the map). I tried to follow the directions on assorted blogs and guides and Tableau help screens, I failed, and after several hours, I had to cut my losses and present something decidedly less than ideal. That’s my visualization story, and I’m sticking to it!
-------------------------------------------------------------------------------------
The final product my classmates will critique is here if you’re curious.