All about Geocoding

Geocoding is how the software draws dots on a digital map. Unfortunately, geocoding is inherently inaccurate. The software is making what amounts to an educated guess by comparing the address of the event to the expected range of addresses along a certain block.

As an example a mall may have numerous shoplifting arrests that show the same address of 6401 S. Main. The mapping software finds a segment of that street that contains an address range of 6401 to 6499 along the east side of the street, and calculates that the point representing that crime should be drawn in the middle of this range. Crimes at 6401 S. Main would all be represented by one dot.

Geocoded crime data, therefore, needs to be considered in light of these limitations. When viewed at a distant scale, geocoded data does a very good job of showing the relative location and density of events. When zoomed in at close range, however, the picture can be deceiving. Geocoded crime points are only approximate. The symbol is not placed exactly where the crime occurred. Dozens of incidents at the same address will look like only one symbol if they are stacked one upon another.

The Tulsa Police Department provides the ability to obtain more detailed data on these stacked crimes by making that crime active and using the rectangle tool to provide details on how many crimes that particular dot represents. See the Crime Mapping Tutorial for specific instructions.

Complicating matters, for a variety of reasons not all crimes can be geocoded. Sometimes the exact location of a crime is unknown. For example, an officer reverses two digits in an address, or a clerk inadvertently adds an extra zero during computer entry. Huge data files describing street segments, although highly accurate, also have their own errors and glitches. The computer software cannot recognize certain addresses that actually exist in the real world but are not present in the street database.