mcpField
S
statfin-mcp

statfin-mcp

Local
github.com/3121nLegal & Government0Updated 20w ago
Repository

About statfin-mcp

statfin-mcp provides MCP access to Statistics Finland’s StatFin tables through PxWeb. It supports table discovery, metadata inspection, and filtered JSON-stat2 data retrieval without an API key.

Key features

  1. 01Search StatFin tables by keyword
  2. 02Inspect table dimensions and value codes
  3. 03Fetch filtered JSON-stat2 data
  4. 04Browse subject areas

Use cases

  1. 01Find tables for a statistical topic
  2. 02Check available filters before querying
  3. 03Retrieve recent figures for selected regions

MCP support

Exposes StatFin table search, metadata, subject browsing, and data retrieval as tools for MCP clients.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "3121n-statfin": {
      "command": "npx",
      "args": [
        "-y",
        "@nor-data/statfin-mcp"
      ]
    }
  }
}
Run commandnpx -y @nor-data/statfin-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@nor-data/statfin-mcp0.1.0

Tools

4

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
emnestringNoOptional: restrict to a subject id (from list_emner), e.g. 'vaerak'
limitintegerNoMax hits (default 10)
querystringYesSearch term, e.g. 'population', 'unemployment', 'consumer price index', 'building'

Details

From the project README

@nor-data/statfin-mcp

MCP server for official statistics from Statistics Finland (Tilastokeskus) — the StatFin database, exposed through the PxWeb API. Search 3000+ tables, inspect their dimensions, and pull data as JSON-stat2. No API key required.

Part of the nor-data family of Nordic open-data MCP servers. Built on the same pattern as the SSB (Norway) server — StatFin and SSB both speak PxWeb, so the tools mirror each other.

Tools

ToolWhat it does
sok_tabellKeyword search across all StatFin tables. PxWeb has no server-side search, so this matches a bundled index (built from the full table tree) locally and returns the most relevant tables with their tabell_id.
hent_tabellMetadata for a table: title + dimensions (variables) with their valid value codes and labels. Read this first to learn which filter codes exist.
hent_dataFetch actual values as JSON-stat2, filtered by a flat {variabelkode: [verdier]} map. Supports special selections like {filter:'top', values:['1']} for the latest period.
list_emnerList the ~140 StatFin subject areas (e.g. vaerak = Population structure). Scope a search or browse when a keyword misses.

All tools accept an optional lang (en default, fi, sv). Variable codes are language-independent; only titles/labels are localised.

Example agent flow

  1. sok_tabell({ query: "population by region" }) → tabell_id: "vaerak/statfin_vaerak_pxt_11ra.px"
  2. hent_tabell({ tabell_id }) → sees variables Alue (area), Vuosi (year), Tiedot (info)
  3. hent_data({ tabell_id, filtre: { Alue: ["SSS"], Vuosi: { filter: "top", values: ["1"] } } }) → latest national figures as JSON-stat2

Install

// Claude Desktop / Claude Code MCP config
{
  "mcpServers": {
    "statfin": {
      "command": "npx",
      "args": ["-y", "@nor-data/statfin-mcp"]
    }
  }
}

Develop

npm install
npm run build-index   # crawl the StatFin tree → data/tables-index.json (~1–2 min)
npm run dev           # run from source via tsx
npm run build         # compile to dist/

The search index is bundled in the package (data/tables-index.json). Re-run npm run build-index to refresh it (nightly is plenty — StatFin updates tables, not the tree, frequently).

Notes & limits
  • PxWeb allows 30 calls / 10 s and rejects queries over ~100 000 cells (HTTP 403) — keep hent_data selections small (few regions × few years).
  • A few subjects (vtp, kivih, sekn, akay) return HTTP 400 server-side and are skipped during indexing; this is a StatFin quirk, not a bug here. The index records them under skipped.
  • Data licensed under CC BY 4.0 by Statistics Finland.

License

MIT

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.