Chart maker
Paste table data to line / bar / pie / scatter charts · export PNG / SVG
一张表进，四种图出
表格的读法很直白。第一列是分类（横轴标签），剩下每一列是一个系列（一条线/一组柱子/一个饼图份额）；表头行自动识别——首行全是数字就当数据行，否则当系列名。数字里带千分位逗号、百分号都能解析；解析不了的格子会明确告诉你"第几行第几列"，而不是悄悄当 0 处理。
每种图有自己的规矩。饼图只取第一个系列（饼图没法表达多系列），出现负值会直接拒绝并说明原因；折线/柱状里解析失败的点按空值留白，图不断裂也不假装；散点图把第一列当 X、各系列当 Y。这些规则全部写死在解析核心里，浏览器与自动化测试跑的是同一份代码。
导出的是真产物。PNG 按 2 倍清晰度渲染，直接可用于文档与群发；SVG 是矢量，放大不糊，适合放进网页或再加工。还能复制完整的 ECharts 配置 JSON——粘进你自己的项目里，就是同样的图。
FAQ
Paste CSV / TSV or text split by spaces or semicolons: the first column is the category (labels), each remaining column is one series; a header row is recommended (it becomes the series names), and a first row that is all numbers is taken as data automatically. Numbers may carry thousands commas and a percent sign.
Line (trends), bar (comparisons), pie (shares - it takes only the first series and refuses negative values), scatter (distributions). After rendering you can copy the complete ECharts configuration JSON to paste into your own project and keep working on it.
Input caps at 200KB, 50,000 data points and 10 series; go past any of them and the page names exactly which one - sample your data and try again. The limits exist to keep the browser from locking up, not to hobble the feature.
Yes. PNG at 2x sharpness and SVG (vector, scales losslessly), plus the configuration JSON copy; the file name is built as original-input-name plus chart type.
No. Parsing, computing and rendering all finish inside the browser (the self-hosted ECharts); the data never leaves the machine, nothing is written to localStorage, reloading clears everything, and the page has no upload endpoint.