示例 CSV
| Warehouse | DeliveryDays |
|---|---|
| Chicago | 2 |
| Chicago | 3 |
| Chicago | 2 |
| Chicago | 4 |
| Chicago | 3 |
| Chicago | 9 |
| Denver | 4 |
| Denver | 5 |
| Denver | 4 |
| Denver | 6 |
| Denver | 5 |
| Denver | 4 |
| Atlanta | 3 |
| Atlanta | 3 |
| Atlanta | 4 |
| Atlanta | 3 |
| Atlanta | 11 |
| Atlanta | 4 |
How to Create a Box Plot
A step by step walkthrough for turning a spreadsheet of raw observations into a published box plot in PlotSet, from shaping the rows to exporting or embedding the finished chart.
约 8 分钟Raw rows in, ranked boxes out: that is how to create a box plot in PlotSet, ending with a published, interactive chart with outliers labelled on purpose and a caption that states what the whiskers mean. Not a demo with placeholder numbers: your own groups and your own observations, run through the same click path a real PlotSet chart takes from upload to publish.
If your spreadsheet already has one row per observation, a group label and a number, this takes about eight minutes. Add ten more if your source is a report someone already summarized into averages or quartiles: you will need to go back for the underlying rows, because a box plot built from a summary is not a box plot, it is a shape drawn to look like one.
你需要的数据
One row per observation, not per group: a group label in the first column, a single number in the second. No pre-calculated quartiles, medians or five-number summaries. PlotSet computes those from the raw rows, so feeding it a summary skips the part that makes the chart honest.
Do not upload a table of quartiles or medians as if it were your data. PlotSet needs the individual observations to compute the five-number summary itself; handing it a summary of a summary produces a median that is wrong more often than it is right.
Keep group labels spelled and capitalized the same way across every row. "Chicago" and "chicago " with a trailing space read as two different groups and split into two boxes.
Give each group enough rows to mean something. A box built from three or four values implies a precision the data does not have; aim for at least ten observations per group where you can.
Extra columns beyond the group and value pair are ignored, so it is fine to keep a date or an ID column in the sheet for your own reference.
操作步骤
PlotSet 中从空白模板到完成图表的具体点击路径。
- 01
Open the box plot template
From the PlotSet dashboard, start a new project and choose the Box Plot template (vertical by default, with a horizontal version available for long group names). It loads with sample data already in place, so you can see the raw, one-row-per-observation shape PlotSet expects before you touch your own file.
每份指南都从这里开始：先选分类，再选模板，不管你要做哪种图表，用的都是同一个网格。 - 02
Shape your sheet into raw rows
Put the group label in the first column and one numeric value in the second, one row per observation. If your source is a pivot table that already shows averages or quartiles per group, go back to the underlying data: that is the input this chart needs, not the summary someone else built from it.
- 03
Import your file or connect a sheet
Use the data panel to import a CSV or .xlsx file, paste rows directly from a spreadsheet, or connect a Google Sheet if you want the chart to refresh when new observations come in. PlotSet reads the first row as headers automatically.
这里的数据视图取自一个条形图项目。你可以上传 CSV、用「Magic Import」从图片里提取数字，或同步 Google Sheet，每种图表类型的面板都一样。 - 04
Map the group and value columns
Confirm which column feeds the group axis and which one holds the numeric value PlotSet will summarize into quartiles. This is usually correct on its own when your headers are clear, so check it manually only if your sheet carries extra columns that could confuse the guess.
绑定决定哪一列作标签、哪一列提供数值、哪一列决定颜色。「Auto Binding」（自动绑定）会根据表头推测，所以要核对一遍，别想当然。 - 05
Sort the boxes by median
Switch the default sort away from alphabetical and order the groups by median instead, ascending or descending depending on the story. Once the boxes are ordered by median, the comparison you built the chart for is visible without the reader having to scan for it.
- 06
Decide what to do with the outliers
Leave the individual points beyond the whiskers turned on and label the ones worth naming. If a group is carrying more outliers than the rest, that is usually a finding, not noise to switch off before anyone notices it.
Box Plot 的「Preview」标签页。所有样式设置都在红框标出的这个面板里，每个设置各自占一个标签页：外观、标题、脚注，以及这种图表类型特有的选项。 - 07
Set the whisker rule and note it in a caption
PlotSet's default draws whiskers to 1.5 times the interquartile range, which covers most cases. If you switch it to min and max instead, say so in a caption: the two conventions look the same on the chart and mean different things.
- 08
Publish it, embed it, or export a file
Publish for a live, interactive chart whose tooltip shows the full five-number summary on hover, grab the embed code to drop it into a page, or export a PNG, SVG or PDF if you need a static image for a deck or a report.
从已加载示例数据的 Box Plot 模板开始，再替换成你自己的数据。
打开模板发布、嵌入或导出
最后一段路对每种图表都一样：「Share」（分享）把图表变成一个在线的公开链接和一段嵌入代码，「Export」（导出）把它输出成静态文件。
Share：公开链接与嵌入代码
打开嵌入代码和公开链接的开关，图表就发布出去了。链接可以放心发给任何人；嵌入代码以 iframe 的形式放进任意页面，悬停提示和交互效果都会保留。之后修改了图表，点「Recreate」就能把改动同步到链接和嵌入代码两边。
Export：SVG、PNG 或 JPG
「Export」会按你需要的尺寸输出一份静态副本。SVG 放到多大都清晰，还能在设计工具里继续编辑；幻灯片和文档用 PNG 或 JPG 最稳妥。想让图片在高分屏上够锐利，关键就是那个倍数设置。
Box Plot 的样式提示
State the whisker rule in the caption
A box drawn to 1.5 times the IQR and one drawn to the minimum and maximum look identical but tell different stories about how extreme the outliers are. One sentence in the caption settles it for the reader instead of leaving them to guess.
Do not force the value axis to start at zero
Unlike a bar chart, a box plot is not encoding a length from a baseline, it is showing where a distribution sits. Forcing the axis to zero on a chart of, say, delivery days in the single digits just compresses every box into a sliver near the top.
Cap it around fifteen boxes on one axis
A handful of distributions read easily side by side; past roughly fifteen, the boxes get thin and the comparison stops working. Filter to the groups the story actually needs and mention in the caption that it is a subset.
Label outliers instead of deleting them
An outlier is usually the most informative point on the chart, not an error. Name the specific one that matters (a date, a location, an ID) rather than hiding it to make the chart look tidier.
制作 Box Plot 时的常见错误
Every box looks like a thin line instead of a box
This is what happens when the sheet holds pre-calculated quartiles or averages instead of raw observations. PlotSet needs the individual values to compute a five-number summary; go back to the source data and upload the actual rows, one per observation.
One box has a whisker stretched far past the others
Check the raw values for that group for a data entry slip, a value in the wrong unit, or simply too few rows. A group with four or five observations produces a box that reacts wildly to a single unusual value.
The boxes sit in alphabetical order, not ranked by median
This means the sort setting is still on its default. Open the sort control in the chart settings and switch it from group name to the median of the value column.
深入了解 Box Plot
不确定 Box Plot 是否是正确选择？查看它是什么、何时使用以及何时不使用。
Box Plot：常见问题
Raw observations, not a summary: one row per data point, a group label in the first column and a single number in the second. PlotSet calculates the quartiles, median and outliers itself from those rows, so a table of pre-computed statistics will not work as the input.
现在制作你的 Box Plot
免费开始，300+ 模板任选。无需信用卡。