Chart2Text

该数据集包含了8,305张图表及其对应的人工撰写摘要，这些图表信息来源于教科书、研究论文和新闻文章等多种渠道。数据集涵盖了柱状图和折线图两种类型，摘要平均由5.2个句子组成，共计113.4个词汇。此外，数据表平均含有32.3个单元格。该数据集的任务旨在进行图表摘要以及自然语言描述的生成。

This dataset comprises 8,305 charts paired with manually authored summaries, which are collected from diverse sources including textbooks, research papers, and news articles. The dataset covers two chart types: bar charts and line charts. On average, each summary consists of 5.2 sentences and contains 113.4 words. Additionally, the average number of cells per data table is 32.3. The task of this dataset focuses on chart summarization and natural language description generation.