Every January, a familiar resolution makes the rounds: “This year, I’m going to read more novels.” You buy a couple of ambitious titles, maybe even join a book club, and for a few weeks it feels great, until life speeds up and the plot starts slipping. Who was that character who vanished for five chapters, and what was their motivation? The challenge lies in the nature of the format: a novel is a massive, unstructured dataset of human behavior that is difficult to consider all at once. But what if you could “see” the entire structure of a story at a glance, mapping the narrative arc just as easily as a sales trend or a supply chain?
In the new paper “Story Ribbons: Reimagining Storyline Visualizations with Large Language Models,” researchers from Harvard University, including D^3 Associate Collaborators Fernanda Viégas and Martin Wattenberg, explore how visualization and LLMs can work together to make stories easier to navigate without flattening their complexity. They introduce Story Ribbons, an interactive system that uses LLMs to extract structured narrative signals—characters, locations, themes, sentiment, and more—from novels and scripts, then turn those signals into customizable storyline visualizations designed to support exploration while helping users calibrate trust in AI-generated insights.
Key Insight: Mapping the “Shape” of a Story
“The issue is how to convert the raw story text into concrete representations of a ‘gentle romance’ or ‘reckless elopement.’” [1]
Story Ribbons starts with a classic visualization idea, each character as a continuous path over time, but treats that form as a flexible interface rather than a fixed diagram. Characters become “ribbons,” where absence creates gaps, thickness can encode narrative importance, and color can even shift to reflect sentiment or custom attributes defined by the user. Crucially, the system supports “explanations on demand,” [2] a feature that allows users to click on any data point and receive a justification from the AI, grounded in the text. This interactivity transforms the visualization from a static chart into a set of dynamic, testable lenses.
Key Insight: AI and Iterative Correction
“Although LLMs proved unreliable at extracting information when used naively, we were able to design a data pipeline that was sufficiently reliable to be helpful to users.” [3]
The research team discovered that simply prompting an LLM to analyze a story produces unpredictable and often flawed results. Their solution: a carefully engineered four-step pipeline that decomposes narrative processing into manageable subtasks, inspired by crowdsourcing workflows. The system first splits novels into chapters, then into scenes based on location changes. For each scene, the AI extracts summaries, conflict and importance ratings, sentiment scores, character lists, and supporting quotes from the text. Critical to the pipeline’s reliability are multiple “correction loops” that catch and fix common LLM errors. When the system discovered that LLMs frequently hallucinate or modify character quotes, it added an exact string matching check. Another correction loop involved deploying a second LLM to consolidate duplicate references, such as when an initial pass fails to recognize that “Jane,” “Jane Bennet,” and “Miss Bennet,” refer to the same person in Pride and Prejudice. Despite these pipelines, Story Ribbons users still identified limitations in the AI’s analytical capabilities, such as its ability to provide holistic analyses that synthesized across chapters. At the same time, users treated the AI as a “partner to bounce ideas off of,” [4] using the tool to challenge their own interpretations, with clear applications in education, such as sparking classroom discussions or helping students find textual evidence for essays.
Why This Matters
For business professionals and executives, this research is less about literature, and more about an AI playbook for turning messy, narrative-heavy domains into decision-grade insights. Strategy decks, customer interviews, incident reports, regulatory filings, even internal emails, all share the same obstacle: they’re rich in meaning but often hard to “see” at scale. Story Ribbons offers a solution: use LLMs to extract structure, produce interactive interfaces and engaging visualizations, and design for trust with explanations and links to original sources.
Bonus
As this article shows, reliability often doesn’t come from a single prompt, but engineered checks, layered methods, and system design. For another look at trust and what it takes to make large systems behave responsibly, check out Teaching Trust: How Small AI Models Can Make Larger Systems More Reliable.
References
[1] Yeh, Catherine et al., “Story Ribbons: Reimagining Storyline Visualizations with Large Language Models,” arXiv preprint arXiv:2508.06772 (2025): 1. https://doi.org/10.48550/arXiv.2508.06772
[2] Yeh et al., “Story Ribbons,” 5.
[3] Yeh et al., “Story Ribbons,” 2.
[4] Yeh et al., “Story Ribbons,” 2.
Meet the Authors
Catherine Yeh is a computer science PhD student at Harvard University.
Tara Menon is Assistant Professor in the Department of English at Harvard University.
Robin Singh Arya is a PhD candidate in the Department of English at Harvard University.
Helen He is a computer science and East Asian studies student at Harvard University.
Moira Weigel is Assistant Professor of Comparative Literature at Harvard University.
Fernanda Viégas is Gordon McKay Professor of Computer Science at the Harvard John A. Paulson School of Engineering and Applied Sciences, and an Associate Collaborator at the Digital Data Design Institute at Harvard (D^3).
Martin Wattenberg is Gordon McKay Professor of Computer Science at the Harvard John A. Paulson School of Engineering and Applied Sciences, and an Associate Collaborator at the Digital Data Design Institute at Harvard (D^3).