关于 data-analysis
data-analysis 是用于分析电子表格和数据文件的技能,面向需要探索数据、计算统计结果并制作可视化的用户。它还能生成公式、查询和结构化分析报告,但不能直接执行代码处理数据。
主要功能
- 01探索数据并计算统计结果
- 02生成洞察与建议
- 03制作图表和可视化
- 04编写公式与查询
- 05设计数据透视表和队列分析
适用场景
- 01分析销售趋势和产品表现
- 02比较客户群体及流失风险
- 03检查预算与实际支出
- 04为数据选择图表形式
安装
复制下面的命令,在终端中运行。
$
npx skills add https:/ / github.com/ claude-office-skills/ skills --skill data-analysis安装后在 Claude Code 中输入/data-analysis确认已加载。
详细介绍
数据分析助手
分析电子表格中的数据,发现洞见,并创建引人入胜的可视化图表。
概览
此技能可帮助你:
- 理解和探索数据
- 执行统计分析
- 生成洞见和建议
- 创建图表和可视化内容
- 编写公式和查询
使用方法
入门
- 分享你的电子表格或数据文件
- 描述你想分析的内容
- 获取洞见、公式或可视化内容
分析类型
探索性分析
"What patterns do you see in this data?"
"Give me an overview of this dataset"
"What are the key statistics?"
具体问题
"What was the total revenue by region?"
"Which products had the highest growth?"
"Is there a correlation between X and Y?"
可视化请求
"Create a chart showing sales trends"
"Make a comparison chart of Q1 vs Q2"
"Show the distribution of customer ages"
输出格式
数据概览
## Dataset Overview
**Rows**: 1,234
**Columns**: 15
**Date Range**: Jan 2025 - Dec 2025
### Column Summary
| Column | Type | Non-null | Unique | Sample Values |
|--------|------|----------|--------|---------------|
| date | Date | 100% | 365 | 2025-01-01 |
| revenue | Number | 98% | 890 | $1,234.56 |
| region | Text | 100% | 5 | North, South |
### Data Quality Issues
- [X] rows have missing values in [column]
- [Y] potential duplicates detected
统计分析
## Statistical Summary
### [Metric Name]
- **Mean**: X
- **Median**: Y
- **Std Dev**: Z
- **Min/Max**: A / B
### Key Findings
1. [Finding with statistical support]
2. [Finding with statistical support]
### Recommendations
- [Action based on analysis]
洞见报告
## Analysis Report: [Topic]
### Executive Summary
[2-3 sentence overview of key findings]
### Key Metrics
| Metric | Value | Change |
|--------|-------|--------|
| Total Revenue | $X | +Y% |
| Avg Order Value | $Z | -W% |
### Trends
1. **[Trend 1]**: [Description with data]
2. **[Trend 2]**: [Description with data]
### Recommendations
1. [Actionable recommendation]
2. [Actionable recommendation]
常见分析工作流
销售分析
1. "Show total sales by month"
2. "Which products are top performers?"
3. "What's the customer segment breakdown?"
4. "Compare this year vs last year"
5. "Forecast next quarter based on trends"
客户分析
1. "What's the customer distribution by segment?"
2. "Calculate customer lifetime value"
3. "Which customers are at risk of churning?"
4. "What's the acquisition cost vs LTV ratio?"
财务分析
1. "Calculate profit margins by product"
2. "What's the expense breakdown?"
3. "Show cash flow trends"
4. "Compare budget vs actual"
公式生成
请求公式
"Write a formula to calculate year-over-year growth"
"Create a VLOOKUP to match customer data"
"Make a dynamic sum based on criteria"
公式输出
## Formula: [Purpose]
### Excel/Google Sheets
```excel
=SUMIFS(Sales[Amount], Sales[Region], "North", Sales[Date], ">="&DATE(2025,1,1))
说明
SUMIFS:对符合多个条件的值求和- 第一个参数:要求和的列
- 后续参数对:条件列 + 条件值
用法
将公式放入你希望显示结果的单元格 [X]。
## Visualization Recommendations
### Choose the Right Chart
| Data Type | Best Chart |
|-----------|------------|
| Trends over time | Line chart |
| Part of whole | Pie/Donut chart |
| Comparison | Bar chart |
| Distribution | Histogram |
| Correlation | Scatter plot |
| Geographic | Map chart |
### Chart Specifications
```markdown
## Recommended Chart: [Type]
**Data Series**:
- X-axis: [Column] (e.g., Date)
- Y-axis: [Column] (e.g., Revenue)
- Series: [Column] (e.g., Region)
**Formatting**:
- Title: "[Descriptive title]"
- Colors: Use consistent color scheme
- Labels: Show values on data points
**Chart Description**:
[What this chart shows and why it's useful]
高级分析
数据透视表设计
## Pivot Table: [Purpose]
**Rows**: [Field 1], [Field 2]
**Columns**: [Field 3]
**Values**: SUM of [Field 4], AVG of [Field 5]
**Filters**: [Field 6]
Expected Output:
| Region | Q1 | Q2 | Q3 | Q4 | Total |
|--------|----|----|----|----|-------|
| North | $X | $X | $X | $X | $X |
| South | $X | $X | $X | $X | $X |
群组分析
## Cohort Analysis
**Cohort Definition**: Customers grouped by [first purchase month]
**Metric**: [Retention rate / Revenue / etc.]
**Time Period**: [12 months]
| Cohort | M0 | M1 | M2 | M3 | ... |
|--------|-----|-----|-----|-----|-----|
| Jan 25 | 100%| 45% | 32% | 28% | ... |
| Feb 25 | 100%| 48% | 35% | 30% | ... |
最佳实践
如何改进分析
- 先清理数据:处理缺失值和重复项
- 明确定义指标:你具体要衡量什么?
- 考虑背景因素:行业基准、季节性
- 验证发现:与其他数据源交叉核对
如何改进可视化效果
- 保持简洁:每张图表传达一个主要信息
- 明确标注:标题、坐标轴、图例
- 使用合适的比例尺:不要通过截断坐标轴造成误导
- 考虑色盲用户:使用图案或易于区分的颜色
局限性
- 无法直接在你的数据上执行代码
- 大型数据集可能需要抽样
- 复杂的统计模型需要专用工具
- 实时数据需要实时连接
- 无法保证对 OCR 识别数据的准确率达到 100%
评论
支持 Markdown