第9讲：OpenClaw 实战：开发数据处理 Skill

动手开发一个完整的数据处理 Skill，掌握 OpenClaw 实战技巧。

一、项目目标

开发「数据分析师」Skill，功能包括：

• 📊 数据清洗和预处理
• 📈 数据统计分析
• 📉 可视化图表生成
• 📄 分析报告导出

二、项目结构

data-analyst-skill/
├── skill.yaml
├── main.py
├── tools/
│   ├── __init__.py
│   ├── data_cleaner.py
│   ├── analyzer.py
│   └── visualizer.py
├── templates/
│   └── report_template.html
├── requirements.txt
└── tests/
    └── test_analyzer.py

三、核心代码

3.1 Skill 主程序

# main.py
from openclaw import Skill
from tools.data_cleaner import DataCleaner
from tools.analyzer import DataAnalyzer
from tools.visualizer import Visualizer

class DataAnalystSkill(Skill):
    def __init__(self):
        super().__init__()
        self.cleaner = DataCleaner()
        self.analyzer = DataAnalyzer()
        self.visualizer = Visualizer()
        
        self.register_intent("analyze_data", self.analyze)
        self.register_intent("clean_data", self.clean)
        self.register_intent("visualize", self.visualize)
    
    def analyze(self, file_path, analysis_type="overview"):
        """数据分析"""
        df = self.cleaner.load(file_path)
        result = self.analyzer.analyze(df, analysis_type)
        return result
    
    def clean(self, file_path, operations=None):
        """数据清洗"""
        df = self.cleaner.load(file_path)
        df = self.cleaner.clean(df, operations)
        return {"cleaned_file": self.cleaner.save(df)}
    
    def visualize(self, file_path, chart_type="bar"):
        """数据可视化"""
        df = self.cleaner.load(file_path)
        chart = self.visualizer.create_chart(df, chart_type)
        return {"chart": chart}

if __name__ == "__main__":
    skill = DataAnalystSkill()
    skill.run()

3.2 数据清洗工具

# tools/data_cleaner.py
import pandas as pd
import numpy as np

class DataCleaner:
    def load(self, file_path):
        """加载数据"""
        if file_path.endswith('.csv'):
            return pd.read_csv(file_path)
        return pd.read_excel(file_path)
    
    def clean(self, df, operations):
        """清洗数据"""
        if 'remove_duplicates' in operations:
            df = df.drop_duplicates()
        
        if 'fill_na' in operations:
            df = df.fillna(df.mean())
        
        if 'remove_outliers' in operations:
            for col in df.select_dtypes(include=[np.number]):
                Q1 = df[col].quantile(0.25)
                Q3 = df[col].quantile(0.75)
                IQR = Q3 - Q1
                df = df[~((df[col] < (Q1 - 1.5 * IQR)) | (df[col] > (Q3 + 1.5 * IQR)))]
        
        return df
    
    def save(self, df, output_path="cleaned_data.xlsx"):
        df.to_excel(output_path, index=False)
        return output_path

四、运行和测试

# 安装依赖
pip install -r requirements.txt

# 运行
openclaw run

# 测试
openclaw test

五、发布到 ClawHub

# 打包
openclaw package

# 发布
openclaw publish --version 1.0.0

---

加入学习群

👉 加入AI编程学习交流群

---

本讲是《Skills 从入门到实践》系列课程的第9讲。

进群

嗨，扫码 进群

微信