Multi-INT Analytics for Pattern Learning and Exploitation (MAPLE) Spatiotemporal Patterns of Life (STPoL) learns normalcy models from observation data. Model clusters capture of where observations usually occur, when they usually occur or how many usually occur. MAPLE's nonparametric machine learning approach learns quickly from limited amounts of unlabeled data and refines models as more data become available. MAPLE models are readily interpretable and allow users to visualize model structure to understand what MAPLE learned. MAPLE exploits these learned models by comparing additional data against them to determine whether the new observations are normal or anomalous. This assessment computes a distance-based metric (e.g., Mahalanobis distance) between an observation and the model. Application of a user-definable threshold makes a decision regarding normalcy of the observation. This automated data analysis allows MAPLE to filter large amounts of data and provide high quality indications and warnings about unusual activity to users. Combining spatial and temporal normalcy produces rich pattern of life models that enable users to answer a range of mission-relevant questions, including the following. Where does a certain activity normally occur? When does that activity normally occur? Is any of this activity currently occurring in an unusual location? Is something not where it should be? Is the activity happening in the usual location, but at an odd time of day?