Mastering .NET Machine Learning

Combining data

Sometimes the data that you obtain from a source system is incomplete. Consider this dataset of crash locations that was obtained from the State Department of Transportation office:

Notice that latitude and longitude are missing and that location does not use a normal address/city/state pattern. Rather, it is OnRoad, Miles, FromRoad, and TowardRoad. Unfortunately, this is fairly common when getting data from public entities—systems may have been built before lat/lon became mainstream and the system's addressing might be designed to only work inside the system. This means we need a way to figure out the latitude and longitude from this ...