Visidata

You can access VisiData functionality through an MCP server that adds powerful data visualization and analysis capabilities. Install, configure, and run the server locally, then use an MCP client to send analytic and visualization requests that produce publication‑ready visuals and in‑depth data insights.

How to use

Connect to the MCP server from your MCP client and use the provided functions to visualize data, analyze skills, and perform core data operations. You can load data from supported formats, generate visualizations like heatmaps and distribution plots, and run advanced analyses such as salary and skills distributions by location. Use the functions to build reproducible data analysis workflows that produce ready-to-share outputs.

Practical usage patterns include: loading a data file, inspecting a sample, performing a full analysis, and generating a visualization or report. You can chain steps to create a complete analysis pipeline, then export results to common formats (CSV, JSON, Excel, images, etc.). The server surfaces a consistent set of functions for visualization, skills analysis, and core data manipulation, so you can build end‑to‑end data workflows with minimal switching between tools.

When you run a local MCP server, your client connects to the local process and issues commands corresponding to the exposed functions. The outputs include generated images, structured data files, and JSON reports that you can save or embed in reports and dashboards.

How to install

Prerequisites you need before installing the MCP server include a working Python 3.10+ environment for the installer and a runtime capable of executing Node‑based packages if you choose the quick install route.

Install options you can use to get started immediately:

# Quick Install (Recommended)
npm install -g @moeloubani/visidata-mcp@beta

# Alternative: Python Install
pip install visidata-mcp

# Development Install
git clone https://github.com/moeloubani/visidata-mcp.git
cd visidata-mcp
pip install -e .

Configuration

Configure your MCP client integrations to know how to reach the VisiData MCP server from your desktop AI assistants or IDEs. The following examples show how to register the server in two common clients.

{ 
  "mcpServers": {
    "visidata": {
      "command": "visidata-mcp"
    }
  }
}

Troubleshooting

If you encounter common setup issues, ensure your MCP server path is correct and your Python version matches the requirements. Restart your client applications after changes and verify that the MCP server process is running in the background.

Verification steps you can run to confirm the server is ready include starting the server and performing a quick readiness check from your client.

# Check if server starts
visidata-mcp

# Test with Python
python -c "from visidata_mcp.server import main; print('✅ Server ready')"

Claude Desktop configuration

Register the MCP server in Claude Desktop so Claude can send requests to VisiData.

{
  "mcpServers": {
    "visidata": {
      "command": "visidata-mcp"
    }
  }
}

Cursor AI configuration

Register the MCP server in your Cursor AI project so you can access VisiData capabilities from your project workflows.

{
  "mcpServers": {
    "visidata": {
      "command": "visidata-mcp"
    }
  }
}

Example usage

Data Visualization examples show how to generate visuals from data files. Use these patterns to produce heatmaps, distributions, and graphs.

# Create a correlation heatmap
create_correlation_heatmap("sales_data.csv", "correlation_heatmap.png")

# Generate distribution plots for all numeric columns
create_distribution_plots("sales_data.csv", "distributions.png", plot_type="histogram")

# Create a scatter plot with categorical grouping
create_graph("sales_data.csv", "price", "sales", "scatter_plot.png", 
            graph_type="scatter", category_column="region")

Skills Analysis and Core Data Tools

You can parse skills, analyze distributions, and perform salary analysis by location and skills using the dedicated functions.

Key operations cover loading data, taking samples, analyzing columns, converting formats, filtering, inspecting statistics, and sorting.

# Parse comma-separated skills into individual columns
parse_skills_column("jobs.csv", "required_skills", "skills_parsed.csv")

# Analyze skills distribution by location
analyze_skills_by_location("jobs.csv", "required_skills", "location", "skills_analysis.json")

# Create skills-location heatmap
create_skills_location_heatmap("jobs.csv", "required_skills", "location", "skills_heatmap.png")

# Comprehensive salary analysis
analyze_salary_by_location_and_skills("jobs.csv", "salary", "location", "required_skills", "salary_analysis.xlsx")

Basic data operations

The server supports loading, sampling, analyzing, converting, filtering, and sorting data across multiple formats.

# Load and analyze data
load_data("data.csv")
get_data_sample("data.csv", 10)
analyze_data("data.csv")

# Transform data
convert_data("data.csv", "data.json")
filter_data("data.csv", "revenue", "greater_than", "1000", "high_revenue.csv")
sort_data("data.csv", "date", False, "sorted_data.csv")

Supported data formats

The MCP server works with a broad range of data formats to fit typical data science and analytics workflows.

Key features and use cases

You can perform complete visualization, advanced skills analysis, location-based metrics, and publication-ready outputs for job market insights, data science workflows, and business intelligence.

Development notes

If you are developing or testing locally, you can install for development, build the package, and run readiness checks.

# Install for development
git clone https://github.com/moeloubani/visidata-mcp.git
cd visidata-mcp
pip install -e .

# Build package
python -m build

# Run tests
python -c "from visidata_mcp.server import main; print('✅ Ready')"

License

MIT License - see LICENSE for details.

Notes

When integrating with client projects, restart your AI application after configuration changes to ensure the MCP server connection is re-established.

Available tools

load_data

Load and inspect data files from various formats.

get_data_sample

Get a preview of your data with configurable row count.

analyze_data

Perform comprehensive data analysis with column types and statistics.

convert_data

Convert between different data formats (CSV ↔ JSON ↔ Excel, etc.).

filter_data

Filter data based on conditions (equals, contains, greater/less than).

get_column_stats

Get detailed statistics for specific columns.

sort_data

Sort data by any column in ascending or descending order.

create_correlation_heatmap

Generate correlation matrices with heatmap visualizations.

create_distribution_plots

Create statistical distribution plots (histogram, box, violin, kde).

create_graph

Custom graphs (scatter, line, bar, histogram) with grouping.

parse_skills_column

Parse comma‑separated skills into individual skills with one‑hot encoding.

analyze_skills_by_location

Skills frequency and distribution analysis by location.

create_skills_location_heatmap

Heatmap showing skills distribution across locations.

analyze_salary_by_location_and_skills

Salary statistics by location and skills combination.