A visual analytics tool for taxonomy-based trajectory data exploration

TL;DR

This paper introduces a visual analytics tool that combines data visualization and machine learning to analyze complex spatio-temporal trajectory data across different domains, enhancing interpretability and insight extraction.

Contribution

The paper presents a novel multi-level approach integrating taxonomy-based classification with visualization for spatio-temporal data analysis, validated through two diverse case studies.

Findings

Successfully classified Arctic fox behaviors and identified distinct movement patterns.

Analyzed hurricane trajectories to reveal shape-influencing angles.

Demonstrated the tool's effectiveness in complex spatio-temporal data analysis.

Abstract

The analysis of spatio-temporal data presents significant challenges due to the complexity and heterogeneity of movement patterns. This project proposes a data analytics tool that combines data visualization and statistical computation to facilitate spatio-temporal data analysis through a multi-level approach. The tool categorizes moving objects into distinct taxonomies using Machine Learning models, adding meaningful structure to the analysis. Two case studies demonstrate the methodology's effectiveness. The first analyzed Arctic fox trajectories, successfully identifying and labeling foxes with Geometric or Kinematic-based behaviors, further categorized into Curvature and Acceleration groups. Statistical indicators revealed that foxes with Acceleration-based behavior showed constant, steady acceleration, while those with Curvature-based behavior exhibited acceleration peaks and sudden deceleration. The second case study examined tropical cyclone data, labeling trajectories with Speed, Curvature, and hybrid Geometric-based behaviors through unique statistical variables. Analysis of hybrid Geometric behavior (Curvature and Indentation combined) identified specific angles with the highest impact on hurricane shape and geometry. The proposed method and tool demonstrate that spatio-temporal data, despite inherent complexity, can be analyzed and explained in detail, providing a theoretical and practical blueprint applicable to multiple domains.