The 24,000 annual lead screening results in Milwaukee, Wisconsin--organized by address--were compared to community demographics and detailed property file records for 250,000 City parcels. The level of detail allowed use of housing characteristics not usually available for risk analysis. ArcView Spatial Analyst was used to introduce perspectives on the clustering of attributes and to avoid use of arbitrary areal units such as ZIP Codes. These tools improved the statistical evaluation of predictors of high-risk neighborhoods. But the methods also raise cautions about the appropriate use of statistical smoothing. The results were organized to support policy efforts by neighborhood and municipal leadership.