LLM/Agent-as-Data-Analyst: A Survey

2025-09-28

AI Summary

Traditional data analysis methods face limitations in semantic understanding, natural language interaction, heterogeneous data processing, and autonomous workflow orchestration. Method: This paper systematically surveys recent advances in large language models (LLMs) and intelligent agent technologies for data analysis, proposing five design principles—semantic awareness, multimodal fusion, autonomous pipeline construction, tool-augmented workflows, and open-task support—to architect an intelligent analytical agent for cross-modal data (structured, semi-structured, unstructured, and heterogeneous). Core techniques include NL2GQL translation, chart and document understanding, cross-modal alignment, and scalable tool invocation mechanisms. Contribution/Results: The study clarifies the technical evolution and application paradigms of LLM-driven data analysis, identifies critical challenges—including interpretability, robustness, and domain adaptation—and outlines future research directions. It provides a systematic foundation for both theoretical advancement and engineering practice in intelligent data analytics systems.

Technology Category

• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Data Mining & Knowledge Management: Intelligent Query Processing

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Large language model (LLM) and agent techniques for data analysis (a.k.a LLM/Agent-as-Data-Analyst) have demonstrated substantial impact in both academica and industry. In comparison with traditional rule or small-model based approaches, (agentic) LLMs enable complex data understanding, natural language interfaces, semantic analysis functions, and autonomous pipeline orchestration. The technical evolution further distills five key design goals for intelligent data analysis agents, namely semantic-aware design, modality-hybrid integration, autonomous pipelines, tool-augmented workflows, and support for open-world tasks. From a modality perspective, we review LLM-based techniques for (i) structured data (e.g., table question answering for relational data and NL2GQL for graph data), (ii) semi-structured data (e.g., markup languages understanding and semi-structured table modeling), (iii) unstructured data (e.g., chart understanding, document understanding, programming languages vulnerable detection), and (iv) heterogeneous data (e.g., data retrieval and modality alignment for data lakes). Finally, we outline the remaining challenges and propose several insights and practical directions for advancing LLM/Agent-powered data analysis.

Problem

• Surveying LLM and agent techniques for data analysis
• Reviewing semantic-aware multimodal data processing methods
• Addressing autonomous pipeline orchestration in data workflows

Innovation

• LLM agents enable autonomous data analysis pipelines
• Tool-augmented workflows enhance semantic analysis functions
• Multimodal integration handles structured and unstructured data