Normalizing data

There is a well-known saying in the data analytics world that “90% of the work is normalizing the data.” There is also a well-known saying at Amazon that “everything fails all the time” (Verner Vogels). This is also true for geospatial data sources. No matter what the data source, there is always the possibility of missing or inaccurate data. It’s common on devices that need high precision to deploy three or more sensors from different manufacturers, all measuring the same attribute, and to then compare the readings against each other. Ideally, all of the sensors will report similar readings, but it’s much easier to identify whether one of the sensors needs to be recalibrated if two of the readings are significantly different from the third. Municipal boundary polygons you may find from public sources can vary widely in terms of spatial accuracy, depending on who digitized them. Incorrect or misaligned coordinates or postal...