ViStruct: Simulating Expert-Like Reasoning Through Task Decomposition and Visual Attention Cues

TL;DR

ViStruct is an automated system that mimics expert reasoning in data visualization by decomposing complex questions into steps and highlighting relevant chart regions, aiding interpretability and future visual literacy tools.

Contribution

It introduces a novel pipeline combining language and vision models to simulate expert-like reasoning in visual analysis tasks.

Findings

Successfully decomposes complex visual questions into structured steps.

Highlights key chart regions aligned with expert reasoning.

Validated with visualization users showing interpretability and alignment.

Abstract

Data visualization tasks often require multi-step reasoning, and the interpretive strategies experts use, such as decomposing complex goals into smaller subtasks and selectively attending to key chart regions are rarely made explicit. ViStruct is an automated pipeline that simulates these expert behaviours by breaking high-level visual questions into structured analytic steps and highlighting semantically relevant chart areas. Leveraging large language and vision-language models, ViStruct identifies chart components, maps subtasks to spatial regions, and presents visual attention cues to externalize expert-like reasoning flows. While not designed for direct novice instruction, ViStruct provides a replicable model of expert interpretation that can inform the development of future visual literacy tools. We evaluate the system on 45 tasks across 12 chart types and validate its outputs with trained visualization users, confirming its ability to produce interpretable and expert-aligned reasoning sequences.