Skip to main content

Usage

Learn how to use MCP Jupyter effectively with your AI assistant.

Basic Usage​

Creating a New Notebook​

Ask your AI assistant to create a notebook:

"Create a new notebook called data_analysis.ipynb"

The AI will:

  1. Create the notebook file
  2. Start a kernel
  3. Be ready for your commands

Working with Existing Notebooks​

"Open the notebook experiments/model_training.ipynb"

Your AI assistant will connect to the existing notebook and preserve all current state.

Key Features​

State Preservation​

All variables, data, and models remain available throughout your session. Work with large datasets without reloading, and hand off complex objects between you and the AI.

Automatic Error Recovery​

The AI sees execution errors in real-time and can automatically install missing packages, fix syntax issues, or suggest corrections.

Seamless Collaboration​

Switch between manual exploration and AI assistance at any point. The AI builds on your work, and you can take over whenever needed.

Smart Package Management​

Missing dependencies are automatically detected and installed, so your workflow isn't interrupted by import errors.

Common Use Cases​

MCP Jupyter excels at collaborative data work. Here are popular use cases:

Data Analysis & Exploration​

  • Data cleaning & profiling: "Handle missing values, outliers, and analyze data quality"
  • Exploratory analysis: "Show me key patterns, distributions, and statistical summaries"
  • Trend analysis: "Plot time series trends with seasonality and correlations"

Machine Learning & Modeling​

  • End-to-end ML pipeline: "Prepare data, engineer features, and compare multiple algorithms"
  • Model optimization: "Tune hyperparameters and evaluate performance comprehensively"
  • Experiment analysis: "Analyze A/B tests and statistical significance"

Data Visualization & Reporting​

  • Automated visualization: "Create appropriate charts and statistical plots for this data"
  • Custom dashboards: "Build interactive visualizations and reports"
  • Anomaly detection: "Identify and visualize unusual patterns"

Research & Advanced Analysis​

  • Hypothesis testing: "Test statistical differences and relationships between variables"
  • Cohort & behavioral analysis: "Track user patterns and segment analysis over time"
  • Concept exploration: "Demonstrate and compare different analytical methods"

Workflow Automation​

  • Data pipelines: "Create repeatable ETL processes and data validation workflows"
  • Report automation: "Generate recurring analysis reports with charts and summaries"
  • Code assistance: "Debug analysis code and explain complex statistical concepts"

Best Practices​

1. Clear Instructions​

Be specific about what you want:

  • ❌ "Analyze the data"
  • ✅ "Perform exploratory data analysis focusing on customer segments and seasonal patterns"

2. Specify Cell Types Clearly​

Help the AI choose the right cell type and operation:

  • For code: "Add a code cell that loads the data"
  • For markdown: "Create a markdown cell with the project title and description"
  • For mixed content: "Add a markdown cell explaining the analysis, then add code to implement it"

3. Handle Operation Errors​

Common AI mistakes and corrections:

  • ❌ AI says "edit_markdown" → ✅ Should be operation="add_markdown" or operation="edit_code"
  • ❌ Putting ASCII art in code cells → ✅ "Put that ASCII art in a markdown cell instead"
  • ❌ IndentationError on non-code content → ✅ "That content belongs in markdown, not code"

4. Iterative Refinement​

Work iteratively with the AI:

1. "Load and preview the customer data"
2. Review the output
3. "Focus on customers from the last quarter" 
4. "Now segment them by purchase frequency"

5. State Management​

  • Keep important variables in the global namespace
  • Use descriptive variable names
  • Periodically check available variables with dir() or locals()

6. Error Recovery​

When errors occur:

  • Let the AI see and handle the error
  • Clarify cell type if there's confusion: "That should be markdown, not code"
  • Provide context if needed
  • The AI will install packages or fix issues automatically

Demo Example​

MCP Jupyter Demo

View the generated notebook →

Tips and Tricks​

  1. Use Markdown cells: Ask the AI to document its analysis
  2. Save checkpoints: Periodically save important state
  3. Combine approaches: Use AI for boilerplate, manually tune details
  4. Leverage errors: Let errors guide package installation
  5. Incremental development: Build complex analyses step by step

Next Steps​