One of the major challenges in health studies with a spatial dimension is to produce valid and meaningful geographical representations of risk. This issue has arisen in our research on childhood asthma and proximity to traffic in Perth, Western Australia. To illustrate the spatial variation in risk over the study area, we developed a method for constructing a "risk field" map and applied this method to our study population. Cases and controls aged 0-19 years were defined using emergency department presentations from 2002 to 2006. For each asthma case, two matched controls were obtained. Geocoded residential addresses were used to calculate "vectors" or arrows of risk across the study area. This allows a rapid interpretation, with the risk of asthma greatest in the direction of the head of the vector relative to the vector's tail. This approach clearly indicated that the risk of asthma presentation at hospital emergency departments is higher for children living closer to the major urban city centers. Application of our method to the study population suggests that the "vector" approach may be useful as an exploratory tool for the spatial investigation of risk of other health outcomes.