Unveiling Motion Patterns through Unsupervised Clustering

In the modern data-driven world, identifying human mobility patterns is crucial. This information aids in urban planning, understanding social networks, disease modeling, transport management, and more. In this study, GPS data were used to track the daily routines of 100 individuals in the city of Kandy, Central Province, Sri Lanka. This type of data and information plays an important role in developing Agent-Based Models(ABMs) that simulate complex social interactions. ABMs have gained popularity in fields such as social sciences, disease propagation, and traffic management. However, the accuracy of ABMs has been hindered by data limitations. With advancements in AI, precise ABMs are becoming essential for practical applications. Existing models often struggle to simulate real-world scenarios accurately. This study aims to address this issue. Through comprehensive mathematical and statistical analyses, it has been possible to identify patterns in social dynamics and subcategories based on occupation, geographical location, and holidays. These insights enhance ABM accuracy, aligning it closely with human perceptual studies. This approach significantly improves ABM fidelity, offering a robust framework for understanding and predicting complex social behaviors.