Data Analyst Cn Zc

数据分析助手 - 数据清洗、统计分析、可视化建议。适合：数据分析师、产品经理、运营。

Security Scan

An instruction-only Chinese data-analysis helper that is internally consistent with its stated purpose (data cleaning, stats, visualization); it requests only python3 and does not ask for credentials, but has minor metadata and dependency-declaration gaps you should be aware of.

Purpose & Capability

The skill's name/description (data cleaning, statistics, visualization) matches the SKILL.md content: pandas-based examples, plotting, time-series, and a report template. However, the skill declares only python3 as a required binary while the instructions assume third-party Python libraries (pandas, matplotlib, seaborn, statsmodels, requests). The metadata in _meta.json (ownerId/slug/version) does not match the registry metadata (different ownerId/slug and SKILL.md version differs from registry version), which is an administrative inconsistency but not necessarily malicious.

Instruction Scope

The instructions are narrowly scoped to reading data (CSV/Excel/JSON/local SQLite/API), cleaning, analysis, plotting, and report generation. There are example calls to requests.get() and sqlite3.connect(), which are expected for data ingestion. The SKILL.md does not instruct the agent to read arbitrary unrelated system configuration or environment variables, nor does it direct data to unexpected external endpoints (the only external example is a placeholder API URL).

Install Mechanism

There is no install spec or code to fetch; this is instruction-only, which minimizes install-time risk. Note: because third-party Python packages are required by the provided examples but not declared, a runtime agent or user might attempt to pip-install packages themselves — the skill does not provide or require an installer but may implicitly rely on package installation.

Credentials

The skill requests no environment variables, no credentials, and no config paths. That is proportionate for a local data analysis assistant. Examples that access an API or a local SQLite DB are normal for the stated functionality; no secrets (API keys/TOKENS) are requested by the skill itself.

Persistence & Privilege

The skill is not set to always: true and does not ask to modify agent-wide configuration. It is user-invocable and allows autonomous model invocation (the platform default), which is expected for skills; no elevated persistence is requested.

Assessment

This skill appears to be what it says: an instruction-only data-analysis helper with example Python code. Before using it: (1) be aware the examples assume Python libraries (pandas, matplotlib/seaborn, statsmodels, requests) that are not declared — confirm these are available or install them in a safe environment; (2) the skill may access files or databases you point it to (e.g., data.csv, database.db) or fetch from URLs — avoid giving it sensitive files or secret-containing DBs unless you trust the execution environment; (3) metadata inconsistencies (ownerId/slug/version) are administrative red flags—verify the source/author if provenance matters; (4) when running generated code, run it in a controlled/isolated environment and back up your data first.

数据分析助手 Skill

快速进行数据清洗、统计分析和可视化。

核心功能

功能	描述
数据清洗	去重、填充、格式化
统计分析	描述统计、相关分析
可视化	图表建议、代码生成
报告生成	自动生成分析报告

使用方法

分析数据

分析这个 CSV 文件：sales.csv

数据清洗

清洗这个数据集，处理缺失值和异常值

生成图表

为这些数据生成折线图代码

Python 数据分析模板

读取数据

import pandas as pd

# CSV
df = pd.read_csv('data.csv')

# Excel
df = pd.read_excel('data.xlsx', sheet_name='Sheet1')

# JSON
df = pd.read_json('data.json')

# 数据库
import sqlite3
conn = sqlite3.connect('database.db')
df = pd.read_sql('SELECT * FROM table', conn)

# API
import requests
response = requests.get('https://api.example.com/data')
df = pd.DataFrame(response.json())

数据预览

# 基本信息
print(df.shape)        # 行列数
print(df.columns)      # 列名
print(df.dtypes)       # 数据类型
print(df.info())       # 详细信息

# 查看数据
print(df.head())       # 前 5 行
print(df.tail())       # 后 5 行
print(df.sample(5))    # 随机 5 行

# 描述统计
print(df.describe())   # 数值列统计
print(df.describe(include='all'))  # 所有列

数据清洗

# 处理缺失值
df.isnull().sum()                    # 统计缺失
df.dropna()                          # 删除缺失行
df.fillna(0)                         # 填充 0
df.fillna(df.mean())                 # 填充均值
df['col'].fillna(df['col'].mode()[0])  # 填充众数

# 处理重复
df.duplicated().sum()                # 统计重复
df.drop_duplicates()                 # 删除重复
df.drop_duplicates(subset=['col'])   # 按列去重

# 数据类型转换
df['date'] = pd.to_datetime(df['date'])
df['price'] = df['price'].astype(float)
df['category'] = df['category'].astype('category')

# 异常值处理
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)]

# 字符串处理
df['name'] = df['name'].str.strip()
df['name'] = df['name'].str.lower()
df['name'] = df['name'].str.replace('old', 'new')

统计分析

# 集中趋势
df['col'].mean()      # 均值
df['col'].median()    # 中位数
df['col'].mode()      # 众数

# 离散程度
df['col'].std()       # 标准差
df['col'].var()       # 方差
df['col'].max() - df['col'].min()  # 极差

# 分布
df['col'].skew()      # 偏度
df['col'].kurt()      # 峰度
df['col'].quantile([0.25, 0.5, 0.75])  # 分位数

# 相关分析
df.corr()             # 相关矩阵
df.corr()['target']   # 与目标的相关性

# 分组统计
df.groupby('category').agg({
    'sales': ['sum', 'mean', 'count'],
    'profit': 'mean'
})

# 交叉表
pd.crosstab(df['col1'], df['col2'])

时间序列分析

# 日期处理
df['date'] = pd.to_datetime(df['date'])
df = df.set_index('date')

# 时间重采样
df.resample('D').sum()      # 按天
df.resample('W').mean()     # 按周
df.resample('M').sum()      # 按月

# 滚动统计
df['rolling_mean'] = df['col'].rolling(window=7).mean()
df['rolling_std'] = df['col'].rolling(window=7).std()

# 时间差
df['diff'] = df['col'].diff()
df['pct_change'] = df['col'].pct_change()

# 季节分解
from statsmodels.tsa.seasonal import seasonal_decompose
result = seasonal_decompose(df['col'], model='additive', period=12)
result.plot()

可视化代码

基础图表

import matplotlib.pyplot as plt
import seaborn as sns

# 设置中文
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False

# 折线图
plt.figure(figsize=(10, 6))
plt.plot(df['date'], df['value'])
plt.title('趋势图')
plt.xlabel('日期')
plt.ylabel('数值')
plt.show()

# 柱状图
plt.bar(df['category'], df['value'])
plt.xticks(rotation=45)
plt.show()

# 散点图
plt.scatter(df['x'], df['y'], alpha=0.5)
plt.xlabel('X')
plt.ylabel('Y')
plt.show()

# 直方图
plt.hist(df['value'], bins=20, edgecolor='black')
plt.xlabel('数值')
plt.ylabel('频数')
plt.show()

# 箱线图
sns.boxplot(data=df, x='category', y='value')
plt.show()

# 热力图
sns.heatmap(df.corr(), annot=True, cmap='coolwarm', center=0)
plt.show()

高级图表

# 分组柱状图
df_grouped = df.groupby(['category', 'type'])['value'].sum().unstack()
df_grouped.plot(kind='bar', figsize=(12, 6))
plt.legend(title='类型')
plt.show()

# 小提琴图
sns.violinplot(data=df, x='category', y='value')
plt.show()

# 配对图
sns.pairplot(df[['col1', 'col2', 'col3', 'category']], hue='category')
plt.show()

# 时间序列
fig, ax = plt.subplots(figsize=(14, 6))
ax.plot(df.index, df['value'], label='实际值')
ax.plot(df.index, df['rolling_mean'], label='7日均值', linestyle='--')
ax.fill_between(df.index, df['lower'], df['upper'], alpha=0.2)
ax.legend()
plt.show()

分析报告模板

def generate_report(df):
    """生成数据分析报告"""
    report = f"""
# 数据分析报告

## 1. 数据概览
- 数据量：{len(df)} 行 × {len(df.columns)} 列
- 时间范围：{df['date'].min()} 至 {df['date'].max()}
- 缺失值：{df.isnull().sum().sum()} 个

## 2. 关键指标
- 总销售额：¥{df['sales'].sum():,.2f}
- 平均订单：¥{df['sales'].mean():,.2f}
- 最高订单：¥{df['sales'].max():,.2f}
- 最低订单：¥{df['sales'].min():,.2f}

## 3. 分布特征
- 偏度：{df['sales'].skew():.2f}
- 峰度：{df['sales'].kurt():.2f}
- 标准差：{df['sales'].std():,.2f}

## 4. Top 5 类别
{df.groupby('category')['sales'].sum().sort_values(ascending=False).head().to_markdown()}

## 5. 趋势分析
- 环比增长：{df['sales'].pct_change().mean()*100:.2f}%
- 月均销售额：¥{df.resample('M', on='date')['sales'].sum().mean():,.2f}

## 6. 建议
1. 重点推广 Top 3 类别
2. 优化低转化品类
3. 关注季节性波动
"""
    return report

注意事项

• 大数据集注意内存使用
• 处理前备份数据
• 结果需要业务验证
• 可视化要简洁清晰

---

创建：2026-03-12 版本：1.0