load-csvLABSCO SUMMARY
A dataset you have not opened yet, explained.
Two tools are enough because the second is a script runner: once the CSV is a dataframe, any analysis is code the assistant writes and executes, and what comes back is the result rather than the code. The prompt template gives it a starting direction so the first question is not yours to invent.
What it is
An MCP server for interactive data exploration. One tool loads a local CSV into a named dataframe, the other runs analysis scripts against it, and a prompt template starts the session from a file path and a topic.
What you get
- A local CSV loaded into a dataframe you name
- Analysis scripts run against the loaded data
- An explore-data prompt template that takes a CSV path and a topic and drives the exploration
- Datasets far larger than a spreadsheet will open, held as a dataframe rather than read into the conversation
Requirements
Python, and a local setup — the repository ships a setup script for macOS. Files are read from local paths, so the data stays on your machine.
Setup effort
Build from source — get the source, then follow the project's own run instructions
USE CASES
01
Getting oriented in a file someone sent you
Load it, then let the exploration run — shape, distributions, and the columns that turn out to matter — before you have decided what to ask.
02
Answering a question that needs a computation
The question is answered by running code over the dataframe, so the figure you get back is computed rather than estimated from a sample.
03
Working a file too large to open in a spreadsheet
Millions of rows stay in the dataframe rather than in a viewer, and only the results of the script cross back into the conversation.
04
Producing a chart or a summary for a report
Ask for the figure you want plotted or tabulated and take the output; the script that made it is reproducible from the same session.
2 tools
Listed in project docs. Your client may show a different list.
run-scriptBEFORE YOU INSTALL
What to know first
Tools you get
2 tools named in the project's documentation.
Requires
uvx on PATH
Where it runs
The documented client runs the server on your machine.
INSTALLATION
Connect your client
Server setup
TERMINAL
python setup.pyMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"mcp-server-ds": {
"command": "uvx",
"args": [
"mcp-server-ds"
]
}
}
}
