This is the dataset the Urban Data Science team manually assembled and used to test the the accuracy of various geocoding services. It contains 40 addresses at three different difficulty levels (for a total of 120 rows). The addresses cover randomly selected residences and businesses across the United States (along with 1 address from Canada). The fields include the full street address, the difficulty level, and the true latitudes/longitudes as confirmed by manually inspecting the center of the buildings.
Data and Resources
|Data Dictionary Files|
|Units of Observation|
|Citation Requirements|| |
Urban Institute. 2020. Revisions. Accessible from https://datacatalog.urban.org/dataset/geocoding-blog-post-data. Developed at the Urban Institute, and made available under the ODC-BY 1.0 Attribution License.