director-data-analysis（总监数据分析）

Data Analysis Skill

Analyze data files (CSV, Excel) and produce actionable insights.

Quick Start

1. Read the file - Use appropriate library:

- CSV: csv module or pandas.read_csv() - Excel: pandas.read_excel() with openpyxl engine

1. Explore the data - Get shape, columns, dtypes, missing values

1. Generate insights - Calculate:

- Descriptive stats (mean, median, mode, std, min, max) - Correlations between numeric columns - Value counts for categorical columns - Trends over time if date column exists

1. Create visualizations - Use matplotlib:

- Bar charts for categorical data - Line charts for time series - Histograms for distributions - Scatter plots for correlations

1. Summarize - Write findings in plain English

Common Patterns

Sales Data

import pandas as pd

df = pd.read_csv('sales.csv')
summary = {
    'total_revenue': df['amount'].sum(),
    'avg_order': df['amount'].mean(),
    'top_products': df['product'].value_counts().head(5),
    'monthly_trend': df.groupby(pd.to_datetime(df['date']).dt.month)['amount'].sum()
}

Customer Data

demographics = df.groupby('segment').agg({
    'age': ['mean', 'median'],
    'income': ['mean', 'std'],
    'id': 'count'
})

Time Series

df['date'] = pd.to_datetime(df['date'])
monthly = df.resample('M', on='date')['value'].sum()

Output Format

Always include:

1. Overview - What the data contains (rows, columns, date range)
2. Key Metrics - Top 5-10 actionable numbers
3. Insights - 3-5 bullet points of what the data reveals
4. Visualizations - At least 2 charts for any dataset with 100+ rows
5. Recommendations - Suggested next steps based on findings

Error Handling

• Handle missing values: df.fillna(0) or df.dropna()
• Handle date parsing: Use pd.to_datetime(..., errors='coerce')
• Handle large files: Process in chunks for files >100MB