Data quality, data precision and data protection

Effective targeting of crime and disorder reduction resources requires,

• good quality data describing these events

• a good understanding of the data that is available for analysis, and

• the application of appropriate types of analyses.

Sticking pins in walls and drawing features on acetates can provide a quick and easy visual solution for describing small numbers of crimes, but does little in supporting an informed, partnered and more strategic approach to directing resources, understanding the drivers behind high levels of crime and disorder, and evaluating the effect of targeted initiatives.

Focused resource targeting requires crime and disorder events to be digitally geocoded and mapped to the precision of the individual property (or location they refer to) or to the relevant full postcode. Geocoding crime and disorder events to this level of precision will have the added benefit of improving the flexibility of referencing this data against other geographic areas and other types of information (see Tackling focus areas for more information about precise geographical referencing). Crime or disorder data that is aggregated to a defined grid network (e.g. a 250m grid square) or a geographic boundary area (e.g. beat or ward) is often too large in size and can hide much of the spatial detail that describes crime and disorder hotspots. Effort should be made to improve the precision and quality of geocoded crime and disorder data to enable partnerships to make more informed decisions for targeting reduction resources (see Data Exchange and Crime Mapping).

Typically, 80% of an analyst’s time that is spent on identifying crime and disorder problem areas is devoted to cleaning the data so that it an be reliably used. This process is most linked to cleaning the address portion of the crime or disorder record so that it can be geocoded precisely. Typical corrections that are required include,

• correcting spelling mistakes in the address fields

• correcting address abbreviations

• correcting unrecognised addresses that are only of local reference to a format that can be recognised

• reformatting records where address information has been entered into the wrong record field.

The geocoding of crime and disorder data is particularly challenging as 35% of this type of data cannot be easily matched to a specific addressable location (e.g. robbery records where the only information that was known was that it occurred on the High Street). This data will requires sanitising in complaiance with Data Protection legislation.

The data quality problem needs to be tackled before partnerships can confidently, precisely, and accurately identify hotspots (see Data Exchange and Crime Mapping [hypertext link to intelligence and information sharing toolkit ] for more details on how to achieve this). Effective software packages are available that provide solutions for virtually automating the cleaning, geocoding, and sanitising of crime and disorder records into a format that can help to effectively target crime and disorder reduction resources.

See the Contacts section of this toolkit for details about these service providers.

Click here for an example  which shows Geocoding crime and disorder records in the London Borough of Brent (May 1997).