# Automated Visualization Code Synthesis via Multi-Path Reasoning and Feedback-Driven Optimization

arXiv: https://arxiv.org/abs/2502.11140 · HTML: https://mlsanity.com/paper/2502.11140/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 14 of its 40 references that are in the map.
- First posted on arXiv 16 Feb 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)

## Builds on (most cited first, 14 of 14)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context](https://mlsanity.com/paper/2403.05530/) (2024)
- [ChartLlama: A Multimodal LLM for Chart Understanding and Generation](https://mlsanity.com/paper/2311.16483/) (2023)
- [MatPlotAgent: Method and Evaluation for LLM-Based Agentic Scientific Data Visualization](https://mlsanity.com/paper/2402.11453/) (2024)
- [Visualization Generation with Large Language Models: An Evaluation](https://mlsanity.com/paper/2401.11255/) (2024)
- [HAIChart: Human and AI Paired Visualization System](https://mlsanity.com/paper/2406.11033/) (2024)
- [LLM4Vis: Explainable Visualization Recommendation using ChatGPT](https://mlsanity.com/paper/2310.07652/) (2023)
- [PlotGen: Multi-Agent LLM-based Scientific Data Visualization via Multimodal Feedback](https://mlsanity.com/paper/2502.00988/) (2025)
- [Prompt4Vis: Prompting Large Language Models with Example Mining and Schema Filtering for Tabular Data Visualization](https://mlsanity.com/paper/2402.07909/) (2024)
- [Foresight: Recommending Visual Insights](https://mlsanity.com/paper/1707.03877/) (2017)
- [NL2INTERFACE: Interactive Visualization Interface Generation from Natural Language Queries](https://mlsanity.com/paper/2209.08834/) (2022)
- [ChartifyText: Automated Chart Generation from Data-Involved Texts via LLM](https://mlsanity.com/paper/2410.14331/) (2024)
- [Chat2VIS: Fine-Tuning Data Visualisations using Multilingual Natural Language Text and Pre-Trained Large Language Models](https://mlsanity.com/paper/2303.14292/) (2023)
- [Automatic Data Visualization Generation from Chinese Natural Language Questions](https://mlsanity.com/paper/2309.07650/) (2023)

## Built on by (most cited first, 7 of 7)

- [PaperBanana: Automating Academic Illustration for AI Scientists](https://mlsanity.com/paper/2601.23265/) (2026)
- [CoDA: Agentic Systems for Collaborative Data Visualization](https://mlsanity.com/paper/2510.03194/) (2025)
- [VisCoder2: Building Multi-Language Visualization Coding Agents](https://mlsanity.com/paper/2510.23642/) (2025)
- [Beyond NL2Code: A Structured Survey of Multimodal Code Intelligence](https://mlsanity.com/paper/2606.15932/) (2026)
- [TopoPilot: Reliable Conversational Workflow Automation for Topological Data Analysis and Visualization](https://mlsanity.com/paper/2603.25063/) (2026)
- [DV-World: Benchmarking Data Visualization Agents in Real-World Scenarios](https://mlsanity.com/paper/2604.25914/) (2026)
- [Beyond Pixels: Introspective and Interactive Grounding for Visualization Agents](https://mlsanity.com/paper/2604.21134/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have become a cornerstone for automated visualization code generation, enabling users to create charts through natural language instructions. Despite improvements from techniques like few-shot prompting and query expansion, existing methods often struggle when requests are underspecified in actionable details (e.g., data preprocessing assumptions, solver or library choices, etc.), frequently necessitating manual intervention. To overcome these limitations, we propose VisPath: a Multi-Path Reasoning and Feedback-Driven Optimization Framework for Visualization Code Generation. VisPath handles underspecified queries through structured, multi-stage processing. It begins by using Chain-of-Thought (CoT) prompting to reformulate the initial user input, generating multiple extended queries in parallel to surface alternative plausible concretizations of the request. These queries then generate candidate visualization scripts, which are executed to produce diverse images. By assessing the visual quality and correctness of each output, VisPath generates targeted feedback that is aggregated to synthesize an optimal final result. Extensive experiments on MatPlotBench and Qwen-Agent Code Interpreter Benchmark show that VisPath outperforms state-of-the-art methods, providing a more reliable framework for AI-driven visualization generation.