FONDECYT Grant 11250039 - Interactive Narrative Analytics
This grant is funded by ANID (Agencia Nacional de Investigación y Desarrollo) and the full title is Interactive Narrative Analytics: Developing scalable knowledge-based narrative extraction models and visual analytics systems for sensemaking in complex information landscapes.
Summary
This project aims to advance the state-of-the-art in computational narrative extraction and interactive analytics to support sensemaking from large, complex collections of news articles. The rapid proliferation of online news and the increasing prevalence of misinformation and disinformation have made it ever more challenging and important for analysts and the public to make sense of the news landscape. Building on the PI’s prior work on narrative maps and a theoretical foundation in narrative extraction and interactive visual analytics, this project will develop novel knowledge-informed models and scalable computational architectures to extract and interactively visualize the key entities, events, and storylines across large news corpora.
Conceptual Framework
The Interactive Narrative Analytics framework integrates five core components: computational architectures for scalable processing, visualization approaches for large-scale narrative data, interaction mechanisms enabling human-in-the-loop semantic interactions, knowledge resources for domain integration and enhancement, and evaluation approaches for assessing narrative quality.
Key Publications
Interactive Narrative Analytics: Bridging Computational Narrative Extraction and Human Sensemaking - B. Keith, IEEE Access 2026. This paper formally defines the field of Interactive Narrative Analytics and establishes its theoretical foundations.
Narrative Maps Visualization Tool (NMVT): An interactive narrative analytics system based on the narrative maps framework - B. Keith, SoftwareX 2025. Presents an open-source interactive tool for extracting and visualizing narrative structures from news collections.
LLM-as-a-Judge Approaches as Proxies for Mathematical Coherence in Narrative Extraction - B. Keith, Electronics 2025. Explores using large language models to evaluate narrative coherence.
VLM-as-a-Judge Approaches for Evaluating Visual Narrative Coherence in Historical Photographical Records - B. Keith, C. Meneses, M. Matus, MC. Castro & D. Urrutia, Electronics 2025. Extends the LLM-as-a-Judge approach to visual narratives using vision-language models.
Web Scraping Chilean News Media: A Dataset for Analyzing Social Unrest Coverage (2019–2023) - I. Molina, J. Morales & B. Keith, Data 2025. Presents a dataset for studying news narratives around Chilean social unrest.