There are over 20,000,000 traffic stops conducted in the United States each year, yet there is no prescribed means to record and report those stops. This prevents the contextualization of news-worthy events pursuant to traffic stops within the broader scope of regional and national averages and trends. To help bridge this gap, we use traffic stop records, census records and geospatial data to model the demographics of annual traffic stops at the county level across the United States. Statistical distance measures are then used to quantify differences between population and predicted traffic stop demographics. To interpret these results spatially, a geospatial smoothing procedure was developed using random contiguous regions that reduces noise, mitigates the influence of outliers, and reveals coherent spatial gradients in estimated disparities. The results show that the relationship between traffic stop subject demographics and local population composition varies by race and sex and does not scale uniformly with population size.The models also reveal distinct differences in how the composition of traffic stops by municipal and state police departments align with population demographics, indicating institution-specific dynamics contribute to demographic disparities in traffic enforcement. Together, these contributions provide a foundation for more contextually informed analysis of traffic stop data in the United States.