church to bar ratio, by U.S. county

Church to bar ratio by county from U.S. Census Bureau data:


The brighter the color, the higher the church to bar ratio. Counties missing data necessary for the computation are shown in black.


From the 2011 County Business Patterns data published at, I extracted the number of establishments in each county that have NAICS codes 813110 (places of worship including churches, temples, mosques, synagogues, etc.) and 722410 (bars, taverns, drink-serving nightclubs, etc.). For each county, I then divided the number of NAICS 813110 establishments by the number of NAICS 722410 establishments to get the “church to bar ratio”. Before dividing I added one to each value to prevent division by zero errors. Finally, I partitioned the resulting ratio distribution into nine color groups, and plotted each county’s ratio color on the map. I used log-transformed values when creating the partitions. Counties missing both NAICS codes in the source data are shown in black on the map. Source code implementing these calculations is available at the link given below, to facilitate peer review.

Possible bias in the result

If NAICS code 813110 excludes church facilities with unpaid staff and clergy—i.e., if the Census Bureau does not to consider them “places of industry” since no one is getting paid to work there—the LDS presence in the west will be severely underrepresented. This warrants further review of NAICS measurement methodology before producing future map updates. Reader comments on the appropriateness of NAICS industry codes for this study are very welcome!

Computational notes

I used SAS to match the data necessary for the ratio computation, and Python to plot the map. Source code for these procedures is given on the Badass Data Science wiki at

Post Author: badassdatascience

3 thoughts on “church to bar ratio, by U.S. county

    Aaron Williams

    (June 28, 2013 - 3:47 am)

    Also, I wonder if church sizes vary from county to county. In Colorado there might be those big mega churches – 1 church for every 10,000 church goers. Elsewhere it might be an average of 1 church for every 100 church goers.

      Aaron Williams

      (June 28, 2013 - 3:48 am)

      I guess the same can be said for Mega-Bars! 🙂


      (June 30, 2013 - 2:05 am)

      It might be possible to normalize by number of employees, which is given in the source data, or number of church employees per capita.

Leave a Reply

Your email address will not be published.