Abstract

The Narrative Maps Visualization Tool (NMVT) is an interactive visual analytics system designed to help analysts understand complex narratives from collections of text documents. NMVT leverages graph-based representations to extract and visualize coherent storylines, showing how events connect over time. The system integrates advanced features including document clustering, coherence-based optimization, storyline extraction, and explainable AI components that provide interpretable insights into narrative connections. NMVT supports both directed analysis (connecting specific events) and exploratory analysis (discovering emerging storylines). By enabling analysts to make sense of large document collections, NMVT addresses critical challenges in intelligence analysis, computational journalism, and misinformation research, allowing users to effectively connect the dots between seemingly unrelated events. The system has been successfully demonstrated on news data by extracting coherent narrative structures that capture both main storylines and alternative perspectives. Case studies show that NMVT's semantic interaction capabilities enable analysts to refine narratives based on domain expertise, while the explainable AI components increase trust in the system's outputs. © 2025 The Author