Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

NL2Dashboard: A Lightweight and Controllable Framework for Generating Dashboards with LLMs

Boshen Shi, Kexin Yang, Yuanbo Yang, Guanguang Chang, Ce Chi, Zhendong Wang, Xing Wang, Junlan Feng

TL;DR

NL2Dashboard is proposed, a lightweight framework grounded in the principle of Analysis-Presentation Decoupling that confines the LLM's role to data analysis and intent translation, while offloading visual synthesis to a deterministic rendering engine.

Abstract

While Large Language Models (LLMs) have demonstrated remarkable proficiency in generating standalone charts, synthesizing comprehensive dashboards remains a formidable challenge. Existing end-to-end paradigms, which typically treat dashboard generation as a direct code generation task (e.g., raw HTML), suffer from two fundamental limitations: representation redundancy due to massive tokens spent on visual rendering, and low controllability caused by the entanglement of analytical reasoning and presentation. To address these challenges, we propose NL2Dashboard, a lightweight framework grounded in the principle of Analysis-Presentation Decoupling. We introduce a structured intermediate representation (IR) that encapsulates the dashboard's content, layout, and visual elements. Therefore, it confines the LLM's role to data analysis and intent translation, while offloading visual synthesis to a deterministic rendering engine. Building upon this framework, we develop a multi-agent system in which the IR-driven algorithm is instantiated as a suite of tools. Comprehensive experiments conducted with this system demonstrate that NL2Dashboard significantly outperforms state-of-the-art baselines across diverse domains, achieving superior visual quality, significantly higher token efficiency, and precise controllability in both generation and modification tasks.

NL2Dashboard: A Lightweight and Controllable Framework for Generating Dashboards with LLMs

TL;DR

NL2Dashboard is proposed, a lightweight framework grounded in the principle of Analysis-Presentation Decoupling that confines the LLM's role to data analysis and intent translation, while offloading visual synthesis to a deterministic rendering engine.

Abstract

While Large Language Models (LLMs) have demonstrated remarkable proficiency in generating standalone charts, synthesizing comprehensive dashboards remains a formidable challenge. Existing end-to-end paradigms, which typically treat dashboard generation as a direct code generation task (e.g., raw HTML), suffer from two fundamental limitations: representation redundancy due to massive tokens spent on visual rendering, and low controllability caused by the entanglement of analytical reasoning and presentation. To address these challenges, we propose NL2Dashboard, a lightweight framework grounded in the principle of Analysis-Presentation Decoupling. We introduce a structured intermediate representation (IR) that encapsulates the dashboard's content, layout, and visual elements. Therefore, it confines the LLM's role to data analysis and intent translation, while offloading visual synthesis to a deterministic rendering engine. Building upon this framework, we develop a multi-agent system in which the IR-driven algorithm is instantiated as a suite of tools. Comprehensive experiments conducted with this system demonstrate that NL2Dashboard significantly outperforms state-of-the-art baselines across diverse domains, achieving superior visual quality, significantly higher token efficiency, and precise controllability in both generation and modification tasks.
Paper Structure (23 sections, 1 theorem, 6 equations, 8 figures, 3 tables)

This paper contains 23 sections, 1 theorem, 6 equations, 8 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Task Formulation
  4. Methodology
  5. Dashboard Generation
  6. Dashboard Modification
  7. Implementation with Multi-Agent System
  8. Theoretical Analysis
  9. Experiment
  10. Experimental Settings
  11. Overall Effectiveness
  12. Efficiency Study
  13. Ablation Study
  14. Case Study
  15. Conclusion
  16. ...and 8 more sections

Key Result

Lemma 1

For any estimator $\mathcal{V}$ of $\mathcal{I}$ with error probability $P_e = P(\mathcal{V} \neq \mathcal{I})$, the conditional entropy is bounded by: where $H_b(P_e)$ is the binary entropy function.

Figures (8)

  • Figure 1: NL2Dashboard Framework with IR
  • Figure 2: Example of a balanced sheet from NL2Dashboard (Dark-style Template)
  • Figure 3: NL2Dashboard with a MultiAgent architecture
  • Figure 4: Translate user prompt into modify script
  • Figure 5: Modification SR with difficulty changing
  • ...and 3 more figures

Theorems & Definitions (1)

  • Lemma 1: Fano's Inequality