At PokitDok, our guiding principle is to create products and services that enable The Business Of Health. To enable our customer facing products, like the PokitDok Marketplace, as well as our enterprise APIs, the PokitDok Platform, our data science team gathers data from numerous sources and turns it into valuable insight. Our newest data science hire, Alec Macrae, put together a series of analysis tools almost immediately after he started with us. Specifically, his tools allow the generation of choropleth maps from our data. What's a choropleth? To the Internet!
"choropleth map: A map that uses graded differences in shading or color (...) inside defined areas on the map in order to indicate the average values of some property or quantity in those areas."
These sort of visualizations have worked their way into the public consciousness recently - if you've seen a "red state/blue state" map of political views per geographic region, you were looking at a choropleth! This post will only serve as a quick introduction to the sorts of analysis possible. We'll be exploring this topic further in future posts.
Let's start with a simple example. What percentage of healthcare providers in each state are MDs? It turns out that we can answer this question quickly and intuitively with a choropleth.
The legend at the top right shows that darker colors indicate a higher percentage. So we see that Maryland has the highest percentage of MDs, followed closely by Virginia, Texas, Louisiana, and Alabama. Next, the same graph, but showing the percentage of dental surgeons - DDSes.
California is clearly a good place to maintain your smile. Let's move on to demographics about the providers themselves.
In this next set of maps, we show percentages of male and female providers per state. Here's the men:
And the women:
Female providers seem to have a majority in Northeast and Midwest states, and, curiously, Alaska, while males dominate Utah and Idaho.
The first two examples graphed demographic information about providers. PokitDok is dedicated to price transparency, and the pricing data we've collected and curated can be inspected, much as the provider data above. For example, how much can you expect to pay for a chest CT scan across the country?
And the following is for a more significant procedure - a hip replacement.
These sample maps only scratch the surface of the sort of analysis and visualization that can be performed. If you had access to this sort of information, what kind of questions would you like to ask? Would analysis tools like this help you as a consumer, or a business? Let us know in the comments.
- Snake_Byte #35: Building Custom Slack Bots with Python - Mimic Your Coworkers - August 11, 2017
- Snake_Byte #14: Writing Made Easy - January 6, 2017
- Top 100K PokitDok Providers - November 30, 2015
- Visualizing Medicare Physician Co-Occurrence and Payment Data Geographically - Part 2 - July 14, 2015
- Visualizing Medicare Physician Co-Occurrence & Payment Data Geographically - Part 1 - July 9, 2015
- National Provider Data Revealed - A Choropleth Map Analysis - October 9, 2014