Abstract. The present paper aims at a method for predicting local rates for crimes against women based on the hipothesis that distinct crime rates have impact on each other. It is proposed that data from SSP-RS and IBGE data are integrated and go through further preprocessing, such as cleaning and discretization, before being analyzed by the Apriori algorithm in order to discover association rules. Preliminary results lead to further experimentation to evaluate potential for helping prevent violence against women.