Data Journalist Agent

Abstract

Data tells stories that shape society, and the data journalist's job is to turn raw information into a story that non-expert audiences can understand and trust through to the end. A high-quality news feature routinely takes a newsroom team weeks, including hunting for context, running statistics, choosing an angle, and designing visuals. Recent agents are individually capable at each individual step: automated data-science agents close the analysis loop, while design agents can synthesize beautiful websites. But can an agent serve as a data journalist end to end? We introduce Data Journalist Agent (Data2Story), a multi-agent framework that orchestrates specialised roles into a single virtual newsroom. Data2Story highlights two innovations over prior approaches. (i) Claims are evidence-grounded and reproducible. We introduce an "Inspector", which associates the intermediate results produced by individual roles so that the numbers, angles, and assets are grounded in data, code, or a reference (e.g., an external URL), so that verifiability is built into the fact itself. (ii) Articles are multimodally generative. Rather than defaulting to plain text and static charts, Data2Story reasons about what its readers will want to read and interact with, then deploys multimodal tools so that the article fits both the data topic and the intended audience (e.g., an interactive map with zoom for a geography piece, or an audio clip for a music piece), making the result readable and engaging. We evaluate Data Journalist Agent on 18 articles from diverse topics and publication sources, each paired with the originally published expert-written piece, along three axes: (a) Human–agent angle coverage, measuring the overlap and complementarity of angles between Data2Story and human-authored articles, to characterize what each side covers; (b) Rubric evaluation with a human study across 53 human participants, with the rubric covering visual design, narrative pacing, data transparency, claim-data alignment, and insight value; (c) Computer-use agents as judges: as an automatic proxy for how real-world users navigate and interact with the article, we employ computer-use agents that fully perceive the interface through actions such as clicking and scrolling; and lastly, (d) Reproducibility, where a coding verifier re-executes every statement against the data and checks that the claims are reproducible or can be grounded in a reference. Our experimental results show that Data2Story receives favorable ratings from human participants and agent judges in both rubric-based evaluation and side-by-side preference. Notably, the Inspector substantially improves the transparency of data and methods, making auditability explicit and measurable. We hope this work moves toward a reproducible, auditable, and data-intelligent system.