Geographically weighted regression (GWR) has been a popular tool widely applied in various disciplines to explore spatial nonstationarity for georeferenced data. Such technique, however, typically restricts the analysis on a single outcome variable to reveal its spatial nonstationary pattern explaining with a set of explanatory variables. When it comes to model multiple interrelated response variables, GWR fails to provide sufficient information of the data as it only allows the separate modeling for each response variable. This study attempts to address the gap by introducing a geographically weighted multivariate multiple regression (GWMMR) technique capable to explore spatial nonstationarity but also to account correlations across multivariate responses. We present the model specification of the proposed method and construct the associated statistical inferences. Certain related modeling issues which include the test of spatial nonstationarity and a semiparametric version of the GWMMR are also discussed. For an empirical illustration, the new technique is applied to the stop-and-frisk data published by the New York Police Department. The analysis results and prediction performance are then compared with those obtained by other existing analytical tools. The results demonstrate the usefulness of the GWMMR in that it can examine the differences possibly overlooked in univariate analysis and understand the multiple outcomes as a system rather than isolated investigations.