标题： LLMs是否准备好帮助非专家用户制作官方统计数据的图表？ 显示英文标题

标题： Are LLMs ready to help non-expert users to make charts of official statistics data?

摘要： 在这个偏见信息、深度伪造和宣传泛滥的时代，可靠数据源的可访问性比以往任何时候都更加重要。 国家统计机构提供经过整理的数据，这些数据包含广泛主题的定量信息。 然而，这些信息通常分散在多个表格中，普通的数字可能难以处理。 因此，这些开放数据可能实际上无法访问。 我们提出问题：“当前的生成式AI模型是否能够帮助识别正确的数据，并完全自动创建图表，以视觉形式提供与用户查询相对应的信息？” 我们对最近大型语言模型（LLMs）的能力进行了结构化评估，评估其根据用户查询从复杂数据生成图表的能力。 与荷兰统计局的多样化公共数据合作，我们评估了多个LLMs在识别相关数据表、执行必要操作以及自主生成适当可视化方面的能力。 我们提出了一种新的评估框架，涵盖三个维度：数据检索与预处理、代码质量以及可视化表示。 结果表明，定位和处理正确数据是最大的挑战。 此外，在没有明确指导的情况下，LLMs很少实施可视化最佳实践。 当提供有关有效图表设计的信息时，模型在表示分数上表现出显著改进。 此外，采用迭代自我评估的代理方法在所有评估维度上都表现出色。 这些发现表明，通过适当的支撑和反馈机制，可以增强LLMs在自动图表生成方面的有效性，并且系统已经能够在三个评估维度上达到必要的准确性。 显示英文摘要

摘要： In this time when biased information, deep fakes, and propaganda proliferate, the accessibility of reliable data sources is more important than ever. National statistical institutes provide curated data that contain quantitative information on a wide range of topics. However, that information is typically spread across many tables and the plain numbers may be arduous to process. Hence, this open data may be practically inaccessible. We ask the question "Are current Generative AI models capable of facilitating the identification of the right data and the fully-automatic creation of charts to provide information in visual form, corresponding to user queries?". We present a structured evaluation of recent large language models' (LLMs) capabilities to generate charts from complex data in response to user queries. Working with diverse public data from Statistics Netherlands, we assessed multiple LLMs on their ability to identify relevant data tables, perform necessary manipulations, and generate appropriate visualizations autonomously. We propose a new evaluation framework spanning three dimensions: data retrieval & pre-processing, code quality, and visual representation. Results indicate that locating and processing the correct data represents the most significant challenge. Additionally, LLMs rarely implement visualization best practices without explicit guidance. When supplemented with information about effective chart design, models showed marked improvement in representation scores. Furthermore, an agentic approach with iterative self-evaluation led to excellent performance across all evaluation dimensions. These findings suggest that LLMs' effectiveness for automated chart generation can be enhanced through appropriate scaffolding and feedback mechanisms, and that systems can already reach the necessary accuracy across the three evaluation dimensions.