Csv Data Summarizer

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from pathlib import Path

def summarize_csv(file_path):
    """
    Comprehensively analyzes a CSV file and generates multiple visualizations.
    
    Args:
        file_path (str): Path to the CSV file
        
    Returns:
        str: Formatted comprehensive analysis of the dataset
    """
    df = pd.read_csv(file_path)
    summary = []
    charts_created = []
    
    # Basic info
    summary.append("=" * 60)
    summary.append("📊 DATA OVERVIEW")
    summary.append("=" * 60)
    summary.append(f"Rows: {df.shape[0]:,} | Columns: {df.shape[1]}")
    summary.append(f"\nColumns: {', '.join(df.columns.tolist())}")
    
    # Data types
    summary.append(f"\n📋 DATA TYPES:")
    for col, dtype in df.dtypes.items():
        summary.append(f"  • {col}: {dtype}")
    
    # Missing data analysis
    missing = df.isnull().sum().sum()
    missing_pct = (missing / (df.shape[0] * df.shape[1])) * 100
    summary.append(f"\n🔍 DATA QUALITY:")
    if missing:
        summary.append(f"Missing values: {missing:,} ({missing_pct:.2f}% of total data)")
        summary.append("Missing by column:")
        for col in df.columns:
            col_missing = df[col].isnull().sum()
            if col_missing > 0:
                col_pct = (col_missing / len(df)) * 100
                summary.append(f"  • {col}: {col_missing:,} ({col_pct:.1f}%)")
    else:
        summary.append("✓ No missing values - dataset is complete!")
    
    # Numeric analysis
    numeric_cols = df.select_dtypes(include='number').columns.tolist()
    if numeric_cols:
        summary.append(f"\n📈 NUMERICAL ANALYSIS:")
        summary.append(str(df[numeric_cols].describe()))
        
        # Correlations if multiple numeric columns
        if len(numeric_cols) > 1:
            summary.append(f"\n🔗 CORRELATIONS:")
            corr_matrix = df[numeric_cols].corr()
            summary.append(str(corr_matrix))
            
            # Create correlation heatmap
            plt.figure(figsize=(10, 8))
            sns.heatmap(corr_matrix, annot=True, cmap='coolwarm', center=0, 
                       square=True, linewidths=1)
            plt.title('Correlation Heatmap')
            plt.tight_layout()
            plt.savefig('correlation_heatmap.png', dpi=150)
            plt.close()
            charts_created.append('correlation_heatmap.png')
    
    # Categorical analysis
    categorical_cols = df.select_dtypes(include=['object']).columns.tolist()
    categorical_cols = [c for c in categorical_cols if 'id' not in c.lower()]
    
    if categorical_cols:
        summary.append(f"\n📊 CATEGORICAL ANALYSIS:")
        for col in categorical_cols[:5]:  # Limit to first 5
            value_counts = df[col].value_counts()
            summary.append(f"\n{col}:")
            for val, count in value_counts.head(10).items():
                pct = (count / len(df)) * 100
                summary.append(f"  • {val}: {count:,} ({pct:.1f}%)")
    
    # Time series analysis
    date_cols = [c for c in df.columns if 'date' in c.lower() or 'time' in c.lower()]
    if date_cols:
        summary.append(f"\n📅 TIME SERIES ANALYSIS:")
        date_col = date_cols[0]
        df[date_col] = pd.to_datetime(df[date_col], errors='coerce')
        
        date_range = df[date_col].max() - df[date_col].min()
        summary.append(f"Date range: {df[date_col].min()} to {df[date_col].max()}")
        summary.append(f"Span: {date_range.days} days")
        
        # Create time-series plots for numeric columns
        if numeric_cols:
            fig, axes = plt.subplots(min(3, len(numeric_cols)), 1, 
                                    figsize=(12, 4 * min(3, len(numeric_cols))))
            if len(numeric_cols) == 1:
                axes = [axes]
            
            for idx, num_col in enumerate(numeric_cols[:3]):
                ax = axes[idx] if len(numeric_cols) > 1 else axes[0]
                daily_data = df.groupby(date_col)[num_col].agg(['mean', 'sum', 'count'])
                daily_data['mean'].plot(ax=ax, label='Average', linewidth=2)
                ax.set_title(f'{num_col} Over Time')
                ax.set_xlabel('Date')
                ax.set_ylabel(num_col)
                ax.legend()
                ax.grid(True, alpha=0.3)
            
            plt.tight_layout()
            plt.savefig('time_series_analysis.png', dpi=150)
            plt.close()
            charts_created.append('time_series_analysis.png')
    
    # Distribution plots for numeric columns
    if numeric_cols:
        n_cols = min(4, len(numeric_cols))
        fig, axes = plt.subplots(2, 2, figsize=(12, 10))
        axes = axes.flatten()
        
        for idx, col in enumerate(numeric_cols[:4]):
            axes[idx].hist(df[col].dropna(), bins=30, edgecolor='black', alpha=0.7)
            axes[idx].set_title(f'Distribution of {col}')
            axes[idx].set_xlabel(col)
            axes[idx].set_ylabel('Frequency')
            axes[idx].grid(True, alpha=0.3)
        
        # Hide unused subplots
        for idx in range(len(numeric_cols[:4]), 4):
            axes[idx].set_visible(False)
        
        plt.tight_layout()
        plt.savefig('distributions.png', dpi=150)
        plt.close()
        charts_created.append('distributions.png')
    
    # Categorical distributions
    if categorical_cols:
        fig, axes = plt.subplots(2, 2, figsize=(14, 10))
        axes = axes.flatten()
        
        for idx, col in enumerate(categorical_cols[:4]):
            value_counts = df[col].value_counts().head(10)
            axes[idx].barh(range(len(value_counts)), value_counts.values)
            axes[idx].set_yticks(range(len(value_counts)))
            axes[idx].set_yticklabels(value_counts.index)
            axes[idx].set_title(f'Top Values in {col}')
            axes[idx].set_xlabel('Count')
            axes[idx].grid(True, alpha=0.3, axis='x')
        
        # Hide unused subplots
        for idx in range(len(categorical_cols[:4]), 4):
            axes[idx].set_visible(False)
        
        plt.tight_layout()
        plt.savefig('categorical_distributions.png', dpi=150)
        plt.close()
        charts_created.append('categorical_distributions.png')
    
    # Summary of visualizations
    if charts_created:
        summary.append(f"\n📊 VISUALIZATIONS CREATED:")
        for chart in charts_created:
            summary.append(f"  ✓ {chart}")
    
    summary.append("\n" + "=" * 60)
    summary.append("✅ COMPREHENSIVE ANALYSIS COMPLETE")
    summary.append("=" * 60)
    
    return "\n".join(summary)


if __name__ == "__main__":
    # Test with sample data
    import sys
    if len(sys.argv) > 1:
        file_path = sys.argv[1]
    else:
        file_path = "resources/sample.csv"
    
    print(summarize_csv(file_path))

pandas>=2.0.0
matplotlib>=3.7.0
seaborn>=0.12.0

{
  "sections": {
    "Notes": "- Automatically detects date columns (columns containing 'date' in name)\r\n- Handles missing data gracefully\r\n- Generates visualizations only when date columns are present\r\n- All numeric columns are included in statistical summary",
    "Files": "- `analyze.py` - Core analysis logic\r\n- `requirements.txt` - Python dependencies\r\n- `resources/sample.csv` - Example dataset for testing\r\n- `resources/README.md` - Additional documentation",
    "⚠️ CRITICAL BEHAVIOR REQUIREMENT ⚠️": "**DO NOT ASK THE USER WHAT THEY WANT TO DO WITH THE DATA.**\r\n**DO NOT OFFER OPTIONS OR CHOICES.**\r\n**DO NOT SAY \"What would you like me to help you with?\"**\r\n**DO NOT LIST POSSIBLE ANALYSES.**\r\n\r\n**IMMEDIATELY AND AUTOMATICALLY:**\r\n1. Run the comprehensive analysis\r\n2. Generate ALL relevant visualizations\r\n3. Present complete results\r\n4. NO questions, NO options, NO waiting for user input\r\n\r\n**THE USER WANTS A FULL ANALYSIS RIGHT AWAY - JUST DO IT.**\r\n\r\n### Automatic Analysis Steps:\r\n\r\n**The skill intelligently adapts to different data types and industries by inspecting the data first, then determining what analyses are most relevant.**\r\n\r\n1. **Load and inspect** the CSV file into pandas DataFrame\r\n2. **Identify data structure** - column types, date columns, numeric columns, categories\r\n3. **Determine relevant analyses** based on what's actually in the data:\r\n   - **Sales/E-commerce data** (order dates, revenue, products): Time-series trends, revenue analysis, product performance\r\n   - **Customer data** (demographics, segments, regions): Distribution analysis, segmentation, geographic patterns\r\n   - **Financial data** (transactions, amounts, dates): Trend analysis, statistical summaries, correlations\r\n   - **Operational data** (timestamps, metrics, status): Time-series, performance metrics, distributions\r\n   - **Survey data** (categorical responses, ratings): Frequency analysis, cross-tabulations, distributions\r\n   - **Generic tabular data**: Adapts based on column types found\r\n\r\n4. **Only create visualizations that make sense** for the specific dataset:\r\n   - Time-series plots ONLY if date/timestamp columns exist\r\n   - Correlation heatmaps ONLY if multiple numeric columns exist\r\n   - Category distributions ONLY if categorical columns exist\r\n   - Histograms for numeric distributions when relevant\r\n   \r\n5. **Generate comprehensive output** automatically including:\r\n   - Data overview (rows, columns, types)\r\n   - Key statistics and metrics relevant to the data type\r\n   - Missing data analysis\r\n   - Multiple relevant visualizations (only those that apply)\r\n   - Actionable insights based on patterns found in THIS specific dataset\r\n   \r\n6. **Present everything** in one complete analysis - no follow-up questions\r\n\r\n**Example adaptations:**\r\n- Healthcare data with patient IDs → Focus on demographics, treatment patterns, temporal trends\r\n- Inventory data with stock levels → Focus on quantity distributions, reorder patterns, SKU analysis  \r\n- Web analytics with timestamps → Focus on traffic patterns, conversion metrics, time-of-day analysis\r\n- Survey responses → Focus on response distributions, demographic breakdowns, sentiment patterns\r\n\r\n### Behavior Guidelines\r\n\r\n✅ **CORRECT APPROACH - SAY THIS:**\r\n- \"I'll analyze this data comprehensively right now.\"\r\n- \"Here's the complete analysis with visualizations:\"\r\n- \"I've identified this as [type] data and generated relevant insights:\"\r\n- Then IMMEDIATELY show the full analysis\r\n\r\n✅ **DO:**\r\n- Immediately run the analysis script\r\n- Generate ALL relevant charts automatically\r\n- Provide complete insights without being asked\r\n- Be thorough and complete in first response\r\n- Act decisively without asking permission\r\n\r\n❌ **NEVER SAY THESE PHRASES:**\r\n- \"What would you like to do with this data?\"\r\n- \"What would you like me to help you with?\"\r\n- \"Here are some common options:\"\r\n- \"Let me know what you'd like help with\"\r\n- \"I can create a comprehensive analysis if you'd like!\"\r\n- Any sentence ending with \"?\" asking for user direction\r\n- Any list of options or choices\r\n- Any conditional \"I can do X if you want\"\r\n\r\n❌ **FORBIDDEN BEHAVIORS:**\r\n- Asking what the user wants\r\n- Listing options for the user to choose from\r\n- Waiting for user direction before analyzing\r\n- Providing partial analysis that requires follow-up\r\n- Describing what you COULD do instead of DOING it\r\n\r\n### Usage\r\n\r\nThe Skill provides a Python function `summarize_csv(file_path)` that:\r\n- Accepts a path to a CSV file\r\n- Returns a comprehensive text summary with statistics\r\n- Generates multiple visualizations automatically based on data structure\r\n\r\n### Example Prompts\r\n\r\n> \"Here's `sales_data.csv`. Can you summarize this file?\"\r\n\r\n> \"Analyze this customer data CSV and show me trends.\"\r\n\r\n> \"What insights can you find in `orders.csv`?\"\r\n\r\n### Example Output\r\n\r\n**Dataset Overview**\r\n- 5,000 rows × 8 columns  \r\n- 3 numeric columns, 1 date column  \r\n\r\n**Summary Statistics**\r\n- Average order value: $58.2  \r\n- Standard deviation: $12.4\r\n- Missing values: 2% (100 cells)\r\n\r\n**Insights**\r\n- Sales show upward trend over time\r\n- Peak activity in Q4\r\n*(Attached: trend plot)*",
    "When to Use This Skill": "Claude should use this Skill whenever the user:\r\n- Uploads or references a CSV file\r\n- Asks to summarize, analyze, or visualize tabular data\r\n- Requests insights from CSV data\r\n- Wants to understand data structure and quality",
    "How It Works": ""
  },
  "content": "This Skill analyzes CSV files and provides comprehensive summaries with statistical insights and visualizations.",
  "id": "csv-data-summarizer-claude-skill_coffeefuelbump",
  "name": "csv-data-summarizer",
  "description": "Analyzes CSV files, generates summary stats, and plots quick visualizations using Python and pandas."
}

---
name: csv-data-summarizer
description: Analyzes CSV files, generates summary stats, and plots quick visualizations using Python and pandas.
metadata:
  version: 2.1.0
  dependencies: python>=3.8, pandas>=2.0.0, matplotlib>=3.7.0, seaborn>=0.12.0
---

# CSV Data Summarizer

This Skill analyzes CSV files and provides comprehensive summaries with statistical insights and visualizations.

## When to Use This Skill

Claude should use this Skill whenever the user:
- Uploads or references a CSV file
- Asks to summarize, analyze, or visualize tabular data
- Requests insights from CSV data
- Wants to understand data structure and quality

## How It Works

## ⚠️ CRITICAL BEHAVIOR REQUIREMENT ⚠️

**DO NOT ASK THE USER WHAT THEY WANT TO DO WITH THE DATA.**
**DO NOT OFFER OPTIONS OR CHOICES.**
**DO NOT SAY "What would you like me to help you with?"**
**DO NOT LIST POSSIBLE ANALYSES.**

**IMMEDIATELY AND AUTOMATICALLY:**
1. Run the comprehensive analysis
2. Generate ALL relevant visualizations
3. Present complete results
4. NO questions, NO options, NO waiting for user input

**THE USER WANTS A FULL ANALYSIS RIGHT AWAY - JUST DO IT.**

### Automatic Analysis Steps:

**The skill intelligently adapts to different data types and industries by inspecting the data first, then determining what analyses are most relevant.**

1. **Load and inspect** the CSV file into pandas DataFrame
2. **Identify data structure** - column types, date columns, numeric columns, categories
3. **Determine relevant analyses** based on what's actually in the data:
   - **Sales/E-commerce data** (order dates, revenue, products): Time-series trends, revenue analysis, product performance
   - **Customer data** (demographics, segments, regions): Distribution analysis, segmentation, geographic patterns
   - **Financial data** (transactions, amounts, dates): Trend analysis, statistical summaries, correlations
   - **Operational data** (timestamps, metrics, status): Time-series, performance metrics, distributions
   - **Survey data** (categorical responses, ratings): Frequency analysis, cross-tabulations, distributions
   - **Generic tabular data**: Adapts based on column types found

4. **Only create visualizations that make sense** for the specific dataset:
   - Time-series plots ONLY if date/timestamp columns exist
   - Correlation heatmaps ONLY if multiple numeric columns exist
   - Category distributions ONLY if categorical columns exist
   - Histograms for numeric distributions when relevant
   
5. **Generate comprehensive output** automatically including:
   - Data overview (rows, columns, types)
   - Key statistics and metrics relevant to the data type
   - Missing data analysis
   - Multiple relevant visualizations (only those that apply)
   - Actionable insights based on patterns found in THIS specific dataset
   
6. **Present everything** in one complete analysis - no follow-up questions

**Example adaptations:**
- Healthcare data with patient IDs → Focus on demographics, treatment patterns, temporal trends
- Inventory data with stock levels → Focus on quantity distributions, reorder patterns, SKU analysis  
- Web analytics with timestamps → Focus on traffic patterns, conversion metrics, time-of-day analysis
- Survey responses → Focus on response distributions, demographic breakdowns, sentiment patterns

### Behavior Guidelines

✅ **CORRECT APPROACH - SAY THIS:**
- "I'll analyze this data comprehensively right now."
- "Here's the complete analysis with visualizations:"
- "I've identified this as [type] data and generated relevant insights:"
- Then IMMEDIATELY show the full analysis

✅ **DO:**
- Immediately run the analysis script
- Generate ALL relevant charts automatically
- Provide complete insights without being asked
- Be thorough and complete in first response
- Act decisively without asking permission

❌ **NEVER SAY THESE PHRASES:**
- "What would you like to do with this data?"
- "What would you like me to help you with?"
- "Here are some common options:"
- "Let me know what you'd like help with"
- "I can create a comprehensive analysis if you'd like!"
- Any sentence ending with "?" asking for user direction
- Any list of options or choices
- Any conditional "I can do X if you want"

❌ **FORBIDDEN BEHAVIORS:**
- Asking what the user wants
- Listing options for the user to choose from
- Waiting for user direction before analyzing
- Providing partial analysis that requires follow-up
- Describing what you COULD do instead of DOING it

### Usage

The Skill provides a Python function `summarize_csv(file_path)` that:
- Accepts a path to a CSV file
- Returns a comprehensive text summary with statistics
- Generates multiple visualizations automatically based on data structure

### Example Prompts

> "Here's `sales_data.csv`. Can you summarize this file?"

> "Analyze this customer data CSV and show me trends."

> "What insights can you find in `orders.csv`?"

### Example Output

**Dataset Overview**
- 5,000 rows × 8 columns  
- 3 numeric columns, 1 date column  

**Summary Statistics**
- Average order value: $58.2  
- Standard deviation: $12.4
- Missing values: 2% (100 cells)

**Insights**
- Sales show upward trend over time
- Peak activity in Q4
*(Attached: trend plot)*

## Files

- `analyze.py` - Core analysis logic
- `requirements.txt` - Python dependencies
- `resources/sample.csv` - Example dataset for testing
- `resources/README.md` - Additional documentation

## Notes

- Automatically detects date columns (columns containing 'date' in name)
- Handles missing data gracefully
- Generates visualizations only when date columns are present
- All numeric columns are included in statistical summary