The Use of Weighted Ternary Histograms for the Visualizationof Segregation

Assessing the level and patterns of residential segregation is an important part of understanding many problems of today's cities. Traditional statistical measures of segregation, such as the exposure indices and the dissimilarity index, are useful but incomplete indicators. This study introduces a new graphical technique, the weighted ternary histogram, which visualizes complex patterns in the location of two or three subgroups of a population. The resultant graphs complement the common indices and expand on their descriptive power in the processes of assessment and hypothesis formulation. When the residential locations of different races in three midsize American cities are compared, the graphs show subtle differences in the pattern of residential segregation among the three cities.