Quick Data

Quick Data is an MCP server that transforms any structured dataset (JSON/CSV) into intelligent, AI-guided analytics workflows.

It intelligently classifies your columns as numerical, categorical, temporal, or identifier types, then recommends the most appropriate analysis methods.

Features

• 🔄 Universal Data Compatibility – Works with any JSON/CSV structure without schema definitions
• 🤖 AI-Powered Analysis Suggestions – Recommends analyses based on detected data characteristics
• 🧮 32 Comprehensive Analytics Tools – From basic statistics to advanced machine learning features
• 📊 Interactive Visualizations – Creates charts and dashboards using Plotly
• 🔍 Automatic Column Detection – Classifies data types and suggests appropriate operations
• 💾 Memory Optimization – Efficient pandas operations with real-time usage monitoring
• 🛠️ Custom Code Execution – Run arbitrary Python analytics code against loaded datasets
• 🎯 Adaptive Conversation Prompts – AI guides users through analytics workflows
• 📈 Multi-Dataset Comparisons – Analyze relationships across different data sources
• ✅ 100% Test Coverage – Production-ready with 130+ passing tests

Use Cases

• Business Intelligence Teams who need to quickly analyze diverse datasets from different departments without writing custom code or defining schemas for each data source
• Data Scientists exploring new datasets who want AI-powered recommendations for initial analysis directions and automated data quality assessments before diving into detailed modeling
• Product Managers analyzing user behavior, sales performance, or market research data who need interactive dashboards and insights without depending on technical teams
• Consultants and Analysts working with client data in various formats who require a universal tool that adapts to any dataset structure while providing professional-grade analytics and visualizations

How to Use It

1. Configure the MCP server for your MCP client:

# Copy the sample configuration
cp .mcp.json.sample .mcp.json
# Find your UV path
which uv
# Example: /Users/yourusername/.local/bin/uv
# Get your project directory
pwd  
# Example: /Users/yourusername/path/to/quick-data-mcp

2. Edit .mcp.json with your actual paths:

{
  "mcpServers": {
    "quick-data": {
      "command": "/path/to/uv",
      "args": [
        "--directory",
        "/path/to/your/quick-data-mcp", 
        "run",
        "python",
        "main.py"
      ],
      "env": {
        "LOG_LEVEL": "INFO"
      }
    }
  }
}

3. Test the server:

uv run python main.py

4. Start with this command to see all available tools:

/quick-data:list_mcp_assets_prompt

Core Analytics Tools

Data Loading & Management

• load_dataset(file_path, dataset_name, sample_size?) – Load any JSON/CSV with automatic schema discovery
• list_loaded_datasets() – Show all datasets with statistics
• clear_dataset(dataset_name) – Remove specific dataset
• get_dataset_info(dataset_name) – Get comprehensive dataset information

Statistical Analysis

• segment_by_column(dataset_name, column_name, method?, top_n?) – Group data by any categorical column
• find_correlations(dataset_name, columns?, threshold?) – Correlation analysis with configurable thresholds
• analyze_distributions(dataset_name, column_name) – Statistical distribution analysis
• detect_outliers(dataset_name, columns?, method) – Outlier detection using IQR or Z-score methods
• time_series_analysis(dataset_name, date_column, value_column, frequency?) – Temporal analysis with trend detection

Advanced Analytics

• validate_data_quality(dataset_name) – Comprehensive quality assessment with 0-100 scoring
• compare_datasets(dataset_a, dataset_b, common_columns?) – Multi-dataset comparison
• merge_datasets(dataset_configs, join_strategy?) – Join datasets with flexible strategies
• calculate_feature_importance(dataset_name, target_column, feature_columns?) – ML feature importance
• execute_custom_analytics_code(dataset_name, python_code) – Execute custom Python code with pandas/numpy/plotly support

Visualization & Export

• create_chart(dataset_name, chart_type, x_column, y_column?, groupby_column?, title?, save_path?) – Generate interactive charts
• generate_dashboard(dataset_name, chart_configs) – Multi-chart dashboards
• export_insights(dataset_name, format?, include_charts?) – Export in JSON, CSV, HTML formats

Dynamic Resources (12 total)

The server provides real-time data context through dynamic resources:

• datasets://loaded – Real-time inventory of loaded datasets
• datasets://{name}/schema – Dynamic schema with column classification
• datasets://{name}/summary – Statistical summary equivalent to pandas.describe()
• analytics://suggested_insights – AI-generated analysis recommendations
• analytics://memory_usage – Real-time memory monitoring

Adaptive Prompts (7 total)

AI-guided conversation starters that adapt to your specific data:

• dataset_first_look(dataset_name) – Personalized exploration guide
• segmentation_workshop(dataset_name) – Interactive segmentation strategy
• correlation_investigation(dataset_name) – Guided correlation analysis
• insight_generation_workshop(dataset_name, business_context?) – Business insight generation

Dynamic Resources (12 total)

The server provides real-time data context through dynamic resources:

• datasets://loaded – Real-time inventory of loaded datasets
• datasets://{name}/schema – Dynamic schema with column classification
• datasets://{name}/summary – Statistical summary equivalent to pandas.describe()
• analytics://suggested_insights – AI-generated analysis recommendations
• analytics://memory_usage – Real-time memory monitoring

Adaptive Prompts (7 total)

AI-guided conversation starters that adapt to your specific data:

• dataset_first_look(dataset_name) – Personalized exploration guide
• segmentation_workshop(dataset_name) – Interactive segmentation strategy
• correlation_investigation(dataset_name) – Guided correlation analysis
• insight_generation_workshop(dataset_name, business_context?) – Business insight generation

Example Workflow

# Load your dataset
await load_dataset("data/sales_data.csv", "sales")
# Get AI-powered first look guidance  
await dataset_first_look("sales")
# Get automatic analysis suggestions
await suggest_analysis("sales")
# Perform suggested analyses
await find_correlations("sales")
await segment_by_column("sales", "customer_segment")
# Create visualizations
await create_chart("sales", "bar", "region", "revenue")
# Assess data quality
await validate_data_quality("sales")

Custom Analytics Code

Execute any Python code against your datasets:

output = await execute_custom_analytics_code("sales", """
# Custom customer segmentation algorithm
customer_scores = df.groupby('customer_id').agg({
    'order_value': ['sum', 'mean', 'count'],
    'date': ['min', 'max']
}).round(2)
# Calculate loyalty scores
customer_scores['loyalty_score'] = (
    customer_scores['total_spent'] * 0.4 +
    customer_scores['order_count'] * 50
).round(1)
print("Analysis completed!")
""")

FAQs

Q: Does this work with any data format or just specific schemas?
A: The server is completely dataset-agnostic. It works with any JSON or CSV file without requiring predefined schemas. The system automatically detects column types and adapts its analysis suggestions accordingly.

Q: What happens if my MCP client doesn’t support resources?
A: The server includes 12 resource mirror tools that provide identical functionality through the tool interface. Tool-only clients can access all resource data using these mirror tools like resource_datasets_loaded().

Q: Can I run custom Python analytics code beyond the built-in tools?
A: Yes, the execute_custom_analytics_code() tool lets you run arbitrary Python code against loaded datasets with full pandas, numpy, and plotly support. This provides unlimited analytical flexibility.

Q: How does the AI analysis suggestion system work?
A: The server analyzes your dataset structure, column types, and data characteristics to recommend appropriate analysis methods. It considers factors like data distribution, column relationships, and missing values to suggest relevant workflows.

Q: What’s the memory usage like for large datasets?
A: The server includes built-in memory optimization with real-time monitoring. You can check memory usage through the analytics://memory_usage resource or memory_optimization_report() tool, and the system provides optimization recommendations.

Q: Can I analyze multiple datasets together?
A: Yes, the server supports multi-dataset workflows through tools like compare_datasets() and merge_datasets(). You can load multiple datasets simultaneously and perform cross-dataset analysis.