Predictive ability for emergency service requests represents a
great potential for a safer community and cost-savings. The ability to predict the
everyday variety of crimes, fires, and emergency medical calls is within reach.
After 9-11, the trend toward information analysis and intelligence in law
enforcement accelerated rapidly. Business intelligence analysis software and
geographic information system technology has found its way into policing, not
just in large urban areas but in small towns as well. National databases and
information sharing among all levels of law enforcement make it possible to
reduce the risk of terrorism threats.
It also works for the crimes that a city such as Richmond, Virginia
experiences routinely. An information management system for predictive crime
analysis includes elements for data mining, reporting, and mapping with GIS
software. Police officers receive the estimations or predictions for crime hot
spots before their shift begins. The result is positive action taken to prevent
crimes rather than a reaction to a crime already committed. Using the system, the
city lowered its dangerous city rating in one year, dropping from fifth highest
to number fifteen. The goal of these systems is to replicate the “intuitive
nature” of a highly experienced police officer. Data collection is the key. Without
baseline data, such systems have no predictive value also critical is a records
management system that facilitates data mining.
While this approach has application for arson crimes, attempting
the same for building fires is unfortunately more problematic. Some progress in
this regard is underway as a team of Australian geographers works with the
Queensland Fire and Rescue Service for the purpose of better allocating fire
service resources and save lives. In the terminology of geographic analysis,
the research team is investigating the spatial-temporal arrangement of urban
fires and their association with weather conditions, calendar events, and
socio-economic conditions. The area protected by this particular fire service
has a large migrant population. The budgets of urban fire-rescue services are
limited and thus essential that managers and planners understand the underlying
forces that drive where, when and why fires start.
Using disaggregated fire incident data form Queensland Fire and
Rescue Service subsequently aggregated to the Statistical Local Area, the team used
the Australian Bureau of Statistics’ defined index of socio-economic
disadvantage (SEIFA) as the basis to identify relationships between
socio-economic disadvantage and building fires. They then used a regression
model to develop predictions for the incidence of building fires over a range
of socio-economic variables.
The geographers identified five significant predictors: percentage
of unemployed, proportion of indigenous population, families living in separate
dwellings, one parent, and parent families with children less than fifteen
years of age. This study shows that mapping urban (building) fires for informed
decision-making and resource allocation has potential for further application
in other areas to validate the results.