The profitability of a dental practice depends greatly on its geographic location. The demographics of the patient population in the area surrounding the practice affects the demand for dental care (that is, the need and the means to pay). Key characteristics are the age and income levels of the population. Another location-dependent factor is the amount of competition from other dental practices. An underserved area could be a good place to start a new practice.
Most dental practices in the United States are privately owned. A sole dentist or partnership will finance the opening of a new dental practice themselves. Placing the practice in a location that will be profitable is an important part of their business strategy.
This is a concern even for dentists who do not own their own practice. Dentists in the US are usually paid as a percent of the amount of dental care they provide, which in turn is dependent of the amount of business the practice can bring in.
In order to find suitable locations for a new dental practice, I compiled demographic data from the US Census Bureau and records of existing practices from online business directories. I confined my search in the states of Kansas and Missouri, with a particular focus on the metropolitan area of Kansas City, Missouri.
All of the code for this project can be found on GitHub.
American Community Survey
I used demographic data from the American Community Survey (ACS) published by the US Census Bureau. This survey contains data from a five-year period, 2012–2016. The attributes collected include age, sex, race, occupation, veteran status, insurance benefits, and many others.
The ACS has some advantages over the Decennial Census. It contains many more attributes than the Census. The data is also aggregated and published annually rather than every ten years, providing a more up-to-date snapshot. One drawback is that the ACS only covers a sample of the population rather than the population entire.
In the ACS, the finest granularity of the geographic component is the Public Use Microdata Area (PUMA). These are generally small geographic regions, containing no fewer than 100,000 people (in order to ensure anonymity and statistical credibility). The boundaries of the PUMAs do not cross state or county lines, so the data can be aggregated at that level. The boundaries also do not cross the Census Blocks and Census Tracts used in the Decennial Census. 
|||They do, however, cross the boundaries of ZIP codes.|
Age and Income
Areas with higher incomes will spend more on dental care than those with lower incomes. This is not surprising, especially in the US where almost all dental care is paid for out-of-pocket or by private insurance.  Areas with older populations will spend more on dental care, because older people tend to have more dental problems and also because of the positive correlation between age and income/wealth. So, we will focus on these two demographic attributes in considering a location.
Below is snippet of the statistics for each PUMA in Kansas (state code 20 ) and Missouri (state code 29).
|State||PUMA||Population||Median Household Income (2016 Dollars)||Median Age||Percent Aged 60 or Older|
I also visualized these results using QGIS.
We see that the highest income areas in these two states are in the suburbs of Kansas City and Saint Louis, on the western and eastern borders of Missouri. There are also slightly higher incomes in the suburbs of Topeka (in north–central Kansas) and Wichita (in south–central Kansas).
The median age is generally higher in more rural regions. In the areas where large universities are located, we see a much lower median age than elsewhere. The youthful patches on the map can be explained by Kansas State University in Manhattan, the University of Kansas in Lawrence, the University of Missouri in Columbia, and Missouri State University in Springfield.
An outlier in income and age is southwest Kansas: it has unusually high incomes and unusually young population for a rural area. My guess is that this is due to oil and gas exploitation in the region, attracting young workers and paying relatively high wages. See the chart from geological survey below.
|||The existing socialized medical programs in the US (e.g., Medicaid and Medicare) generally do not provide dental coverage. One exception is Tricare, which does.|
|||The codes used in the ACS data accord with the Federal Information Processing Standard (FIPS) state codes.|
A third consideration is the number of dental practices already operating the region. The more dental practices there are, the more competition. This leads to lower utilization of the capacity of each dental practice and lower profitability.
To assess this, I scraped the listings of dentists and dental practices from an online business directory. Since I wanted to focus on the metropolitan area of Kansas City, I collected listing from the region around Lenexa, Kansas, an inner-ring suburb. 
The shapefiles provided by the Census Bureau for the PUMAs contain the boundaries of these geographic regions. Using the fiona Python library, we can easily open and manipulate shapefiles. In conjunction with the shapely library, we can find which PUMA each of the dental practices is located in, by way of its coordinates.
After removing duplicate locations from our list of dental practices, we can then tabulate the number in practices in each PUMA. One rule of thumb is that an ideal practice has 2000 active patients. So, we would be looking for areas with around 2000 or more people per practice.
According to our chart, southwest Johnson County (bottom left PUMA) and western Kansas City (two PUMAs on right) have a suitable ratio. Wyandotte County (top center PUMA) is somewhat underserved, so it could be a profitable location for a practice. Southeast Johnson County (bottom center PUMA) has the lowest ratio; however, it is also the area with the highest median household income Kansas or Missouri, so it may nevertheless be a good location.
|||Since I did not scrape the entire business directory, only the data for the areas close to the origin of Lenexa, KS, can be expected to be reasonably complete.|
The attributes considered here may be a good start, but there are other factors worth looking at. Future analysis could investigate the dental insurance providers in the area and how much they will pay for various dental procedures. The same procedure in one area may be better compensated than in another due to the typical insurance coverage of the patients.
The data quality of the existing dental practices could be improved. The addresses scraped from the business directory were quite dirty, having problems like:
- Some dentists are listed multiple times at different practices
- Some dental practices listed are no longer operating
- Errors in addresses for practices, such as the wrong city or ZIP code given
There is also little trust in the listings being complete. In the future, there may be better sources for this information, such as lists of in-network dental practitioners published by insurance companies.
Open Street Map's Nominatim geocoding API seems to be less tolerant to malformed addresses than Google Maps'. For a commercial application, it would probably be worth paying for Google's service.