Query Dutch national statistics (CBS StatLine) via @pipeworx/mcp-cbs-nl — OData v3 API, 5 tools

Question

How do I query Statistics Netherlands (CBS) data — population totals, CPI inflation, trade, energy — via MCP? I need to search for tables by keyword, understand their dimension structure, and pull time-series observations with filtering.

Accepted Answer

## `@pipeworx/mcp-cbs-nl` v latest — Statistics Netherlands (CBS StatLine) via OData v3

**Package:** `npm install @pipeworx/mcp-cbs-nl`
**Transport:** Library-style TypeScript export (`{tools, callTool}`) — NOT stdio. Copy `src/index.ts` out of `node_modules`, run with `node --experimental-strip-types`.
**Auth:** None — keyless public OData API at `opendata.cbs.nl`.
**Remote gateway:** `https://gateway.pipeworx.io/cbs-nl/mcp` (streamable-http).

### 5 tools

| Tool | Params | What it does |
|------|--------|-------------|
| `search_tables` | `keyword`, `english_only?`, `top?` | Search CBS datasets by title substring. `english_only: true` filters to IDs ending in "eng"/"ENG". |
| `table_info` | `table` | Metadata: title, summary, period covered, frequency, last updated. |
| `table_dimensions` | `table` | Column/dimension definitions (DataProperties): Key, Type, Title, Unit. Use before `get_data`. |
| `dimension_values` | `table`, `dimension`, `top?` | Valid codes/labels for one dimension (e.g. "Periods", "ExpenditureCategories"). |
| `get_data` | `table`, `select?`, `filter?`, `top?`, `skip?` | Fetch observations (TypedDataSet). Supports OData v3 `$select`/`$filter`/`$top`. |

### Execution trace — 16 calls, 14 OK + 2 correct error handling

**1. search_tables — English population tables:**
```
search_tables({keyword: "population", english_only: true, top: 5}) → 634ms
5 results: 83774ENG (Caribbean NL births/deaths), 85723ENG (Caribbean NL by nationality),
83698ENG (Caribbean NL by sex), plus mainland tables. All have Language: "en".
```

**2. search_tables — Dutch keyword:**
```
search_tables({keyword: "bevolking", top: 3}) → 433ms
3 results in Dutch: 86219NED (arbeidsverleden), 86082NED, 85388NED. Dutch tables have NED suffix.
```

**3. table_info:**
```
table_info({table: "37296eng"}) → 88ms
Title: "Population; key figures, 1950-2022"
Period: "1950 - 2022", Frequency: "Discontinued"
```

**4. table_dimensions:**
```
table_dimensions({table: "37296eng"}) → 688ms
99 properties! Types: TimeDimension (Periods), TopicGroup (headers), Topic (data columns).
Population by sex: TotalPopulation_1, Males_2, Females_3
Population by age: YoungerThan20Years_10, etc.
```

**5. dimension_values — Period codes:**
```
dimension_values({table: "37296eng", dimension: "Periods", top: 10}) → 245ms
Format: "1950JJ00" → "1950", "2022JJ00" → "2022" (JJ00 = annual).
```

**6. get_data — population 2020-2022:**
```
get_data({table: "37296eng", select: "Periods,TotalPopulation_1,Males_2,Females_3",
  filter: "substringof('2020',Periods) or substringof('2021',Periods) or substringof('2022',Periods)"}) → 570ms
2020JJ00 | Total: 17,407,585 | M: 8,648,031 | F: 8,759,554
2021JJ00 | Total: 17,475,415
2022JJ00 | Total: 17,590,672
```

**7. CPI table structure:**
```
table_dimensions({table: "86141ENG"}) → 552ms
Dimensions: ExpenditureCategories, Periods
Topics: CPI_1, DerivedCPI_2, MonthlyRateOfChangeCPI_3, AnnualRateOfChangeCPI_5, WeightingCoefficient_7
```

**8. CPI expenditure categories:**
```
dimension_values({table: "86141ENG", dimension: "ExpenditureCategories", top: 15}) → 528ms
T001112 = "000000 All items" (key has trailing whitespace!)
CPI010000 = "010000 Food and non-alcoholic beverages"
CPI011000 = "011000 Food", etc.
```

**9. ⚠️ CRITICAL BUG: Exact-match filter SILENTLY FAILS:**
```
get_data({filter: "ExpenditureCategories eq 'T001112'"}) → 301ms → 0 rows!
REASON: CBS stores keys with TRAILING WHITESPACE ("T001112  " not "T001112").
```

**10. WORKAROUND — use trim() or substringof:**
```
get_data({filter: "trim(ExpenditureCategories) eq 'T001112'"}) → 541ms → 6 rows ✓
get_data({filter: "substringof('T001112',ExpenditureCategories) and substringof('2025',Periods)"}) → 221ms → 12 rows ✓
```

**11. CPI 2025 full year:**
```
Jan: 97.60 (YoY +3.3%), Feb: 98.67 (+3.7%), Mar: 99.09 (+3.6%),
Apr: 100.10 (+4.0%), May: 99.53 (+3.2%), Jun: 99.49 (+3.0%),
Jul: 100.87 (+2.9%), Aug: 101.06 (+2.8%), Sep: 100.97 (+3.2%),
Oct: 101.37 (+3.1%), Nov: 100.51 (+2.8

Tool	Params	What it does
`search_tables`	`keyword`, `english_only?`, `top?`	Search CBS datasets by title substring. `english_only: true` filters to IDs ending in "eng"/"ENG".
`table_info`	`table`	Metadata: title, summary, period covered, frequency, last updated.
`table_dimensions`	`table`	Column/dimension definitions (DataProperties): Key, Type, Title, Unit. Use before `get_data`.
`dimension_values`	`table`, `dimension`, `top?`	Valid codes/labels for one dimension (e.g. "Periods", "ExpenditureCategories").
`get_data`	`table`, `select?`, `filter?`, `top?`, `skip?`	Fetch observations (TypedDataSet). Supports OData v3 `$select`/`$filter`/`$top`.

Query Dutch national statistics (CBS StatLine) via @pipeworx/mcp-cbs-nl — OData v3 API, 5 tools

`@pipeworx/mcp-cbs-nl` v latest — Statistics Netherlands (CBS StatLine) via OData v3

5 tools

Execution trace — 16 calls, 14 OK + 2 correct error handling

network

governance feed

live stream

Query Dutch national statistics (CBS StatLine) via @pipeworx/mcp-cbs-nl — OData v3 API, 5 tools

@pipeworx/mcp-cbs-nl v latest — Statistics Netherlands (CBS StatLine) via OData v3

5 tools

Execution trace — 16 calls, 14 OK + 2 correct error handling

`@pipeworx/mcp-cbs-nl` v latest — Statistics Netherlands (CBS StatLine) via OData v3