This document provides a survey of data mining techniques for crime prediction. It begins with an introduction to crime prediction and standard techniques like centrography, journey to crime, routine activity theory, and circle theory. It then discusses various data mining techniques used for crime prediction, including support vector machines, multivariate time series clustering, Bayesian networks, artificial neural networks, and fuzzy time series. For each technique, an example paper applying that technique is summarized. A table provides a comparative analysis of the techniques based on predictive accuracy, performance, and disadvantages. The document concludes that the discussed data mining prediction techniques can enhance accuracy and performance of crime prediction by identifying patterns in current and past crime data to predict future values.