Trajectory pattern mining involves analyzing spatial and temporal location data to identify recurring movement patterns. This technique uses GPS and GSM data to extract meaningful patterns from movement trajectories. The goal is to represent spatial patterns through sequences of locations connected by time intervals. For example, a pattern might be represented as Bistro Square → Appleton Tower → Potterrow → Chambers Str → The Mound → Waverly Station. Possible applications include predicting movement, analyzing aggregate behavior, discovering regions of interest, and identifying traffic flow and blockages. The data structure involves converting raw (latitude, longitude, timestamp) data into (x, y, t) coordinates. T-patterns are sequences of locations with transitions between them, defined by a set of states and transitions. The curse of dimensionality poses challenges due to the high-dimensional space of trajectories. To address this, Regions of Interest (ROIs) are used to reduce the search space. ROIs are static or dynamically determined points of interest, and popular points are identified based on density distribution. T-pattern mining involves two approaches: static and dynamic. The static approach computes ROIs over all trajectories and finds sequences that meet support thresholds. The process iteratively extends patterns by projecting between regions. The results include patterns like A→F→D, A→F→E, B→D, and B→E. The static approach is effective for applications requiring simple, coarse-grained results. Future work includes using T-patterns to predict future movement and extending SQL with TAS queries through the DAEDALUS framework. The research has been cited for its contributions to ROI discovery and trajectory pattern mining.Trajectory pattern mining involves analyzing spatial and temporal location data to identify recurring movement patterns. This technique uses GPS and GSM data to extract meaningful patterns from movement trajectories. The goal is to represent spatial patterns through sequences of locations connected by time intervals. For example, a pattern might be represented as Bistro Square → Appleton Tower → Potterrow → Chambers Str → The Mound → Waverly Station. Possible applications include predicting movement, analyzing aggregate behavior, discovering regions of interest, and identifying traffic flow and blockages. The data structure involves converting raw (latitude, longitude, timestamp) data into (x, y, t) coordinates. T-patterns are sequences of locations with transitions between them, defined by a set of states and transitions. The curse of dimensionality poses challenges due to the high-dimensional space of trajectories. To address this, Regions of Interest (ROIs) are used to reduce the search space. ROIs are static or dynamically determined points of interest, and popular points are identified based on density distribution. T-pattern mining involves two approaches: static and dynamic. The static approach computes ROIs over all trajectories and finds sequences that meet support thresholds. The process iteratively extends patterns by projecting between regions. The results include patterns like A→F→D, A→F→E, B→D, and B→E. The static approach is effective for applications requiring simple, coarse-grained results. Future work includes using T-patterns to predict future movement and extending SQL with TAS queries through the DAEDALUS framework. The research has been cited for its contributions to ROI discovery and trajectory pattern mining.