●Guide 02
为动态图表准备 CSV 数据
一步一步。
Most chart problems start in the source table. A simple, rectangular dataset makes the animation predictable and gives you more time to focus on the story.
At a glance 你将学到什么
从干净的数据到可以发布的图表。
- Use one clear header row
- Keep labels short and unique
- Normalize units and time periods
- Remove totals, notes, and merged cells
一份可直接做图的 CSV 示例一行表头、简短标签、一致的数值
Channel,2023,2024,2025
Search,42,58,71
Email,35,49,63
Social,28,46,59
Partner,22,31,4401 教程
跟随工作流
新图表按顺序做完这些步骤，也可当作已有动画的检查清单。
- 01
Define the grain
Decide what one row represents and keep it consistent. A row might represent a country, product, month, or channel—but never a mixture of levels.
- 02
Create one header row
Use a unique, descriptive name for every column. Put the category field first, followed by time periods or numeric measures.
- 03
Normalize numeric values
Use the same unit and decimal convention throughout each series. Remove currency symbols and thousands separators if they interfere with parsing.
- 04
Standardize time labels
Choose one date interval and format, such as 2022–2026, Q1–Q4, or Jan–Dec. Sort periods chronologically before import.
- 05
Handle missing data intentionally
Confirm whether a blank means zero, unavailable, or not applicable. Do not silently replace missing values unless that treatment is analytically valid.
- 06
Run a final quality check
Scan for duplicate labels, inconsistent units, unexpected totals, footnotes, and formulas that did not export as values.
02 视觉示例
试试这些图表模板
先比较运动和阅读方式，再决定格式。每个示例都是可编辑、由数据驱动的图表。
03 应用
相关场景
04 FAQ
常见问题
Should categories go in rows or columns?
For ranking animations, put each category in a row and each time period in a column. For a simple trend, put time periods in the first column and one or more measures in the remaining columns.
How should percentages be formatted in CSV?
Use a consistent convention across the dataset, such as 42 for 42 percent or 0.42 for a decimal share. Confirm the chart’s display format after import.
What should I do with missing values?
Determine whether the value is truly zero or simply unavailable. Preserve that distinction during cleaning, because replacing unavailable data with zero can create a misleading animation.