Using Rill MCP

Trying Rill MCP implementation with Claude Desktop. It worked on the first try with a local Rill and NYC and Pickup data set.

I think we arrived at Self-Service BI after many years of trying if we roll out MCP capabilities to everyone in the org.

Beyond mere numbers, we get interpreted context, too.

Docs: Rill MCP Server | Rill

Question:

Using MCP Rill, can you give “avg tip” per ride in NYC with pickup zone JFK Airport?

Claude did this:

I’ll help you find the average tip per ride for trips picked up from JFK Airport in NYC using the available metrics data. Let me first explore what metrics views are available and then query for the specific information you need.Now let me examine the structure of these metrics views to understand what data is available:Perfect! I can see there’s an average_tips_measure that gives us the average tip per trip, and a pickup_zone dimension. Let me also check the time range available for this data:Now let me query for the average tip per ride specifically for JFK Airport pickup zones:Based on the data from the NYC trips metrics, here’s what I found:

Average Tip per Ride from JFK Airport: $3.72

Additional context:

• Total trips from JFK Airport: 387,668 trips
• Data period: April 30, 2023 to June 1, 2023

This means that for rides originating from JFK Airport in NYC during this period, passengers tipped an average of approximately $3.72 per trip. The data covers a substantial sample size of nearly 400,000 trips, making this a statistically robust average.

Full image

The dashboard looks like this with the same filters, and we see it’s spot on:

claude_desktop_config.json:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

{
  "mcpServers": {
    "rill": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://localhost:9009/mcp/sse"
      ]
    }
  },
  "globalShortcut": ""
}

Adding in Claude Desktop on MacOS:

See Claude Configs (MCP)

Origin: Rill
References: LLM, GenBI