Risk Terrain Modelling

Dr Nick Malleson

[Fullscreen]

Risk Terrain Modelling - Overview

RTM book cover

GIS techniques to explore the relationship between crime and spatial factors that influence it. E.g.

Public transport

Drug dealers

Schools

Gangs

Parks

Shops

...

Theoretically / empirically informed

Can be used to make predictions

Example: Children cluster

RTM example - school children clustering

Source: RTM introduction webinar

About The Authors

Methods and software developed by the Rutgers Centre on Public Security

Documentation on the website

RTM software

Free for education

Additional features in the enterprise version

Caplan, J.M., Kennedy, L.W., 2010. Risk Terrain Modelling Manual. Rutgers Centre on Public Security, Newark, NJ.

Main Principles

Crime risk is associated with geography

Spatial factors increase/decrease the risk

BUT: relationship between risk factors and geography is non-trivial

Different configurations of risk factors will determine overall risk

"Drawing from an example in meteorology, individual factors that are incorporated into weather forecasting do not necessarily produce rain, thunder storms or hurricanes by themselves. It is only when they intersect in space and time that they have the greatest potential to yield a particular outcome, e.g. a storm. Other times, only one or a few factors may be required to interact in the same geography and at certain times for a particular event to occur." (Caplan and Kennedy, 2010).

The Process

Identify risk factors to test

Theoretically informed

Empirically informed

Decide on their type of influence

density

proximity

And the extent of their spatial influence

Use the RTM software to create a regression model

Crime = α + β Factor1 + β Factor2 + ...

Numerous models are created / tested and the optimal one is returned

Use statistical / GIS software to generate a risk surface

RTM Example: Burglary

output risk surface

Inputs:

Crime incidence data (dependent variable).

Detached houses

Bus Stops

Output:

A model, and a risk surface

Exp(-3.3565 + 3.0617 * "bus stops" + 2.1578 * "det points") / Exp(-3.3565)