statswalesr

statswalesr is an R package for accessing data from the StatsWales public API v2.

Installation

install.packages("statswalesr")

Usage

Discover datasets

library(statswalesr)

# List all published datasets (the full catalogue is fetched automatically)
datasets <- statswales_list_datasets()

# Full-text search
results <- statswales_search("hospital")
results_fuzzy <- statswales_search("transprt", mode = "fuzzy")

# Browse by topic
topics <- statswales_list_topics()
topic_content <- statswales_get_topic(topics$id[1])

Retrieve data

Dataset IDs are UUIDs returned by statswales_list_datasets() or statswales_search().

id <- datasets$id[1]

# Entire dataset, human-readable column names and values (default)
df <- statswales_get_dataset(id)

# A single page of rows, if you don't want everything
df_page <- statswales_get_dataset(id, all_pages = FALSE, page_size = 100)

# Welsh language
df_cy <- statswales_get_dataset(id, lang = "cy-gb")

Results are tidied for analysis by default: the API’s internal *_sort columns are dropped, whitespace padding is stripped, and numeric columns (including the data values) are converted to numbers. Pass tidy = FALSE to get the API response as-is.

Filter data

The filter argument is a list of named lists. Multiple list elements use AND logic; multiple values within one element use OR logic.

# See what dimensions can be filtered
filters <- statswales_get_filters(id)
filters[[1]]$columnName   # dimension name (use as the key)
filters[[1]]$values       # data frame of reference codes and labels

# Filter to specific values
df_filtered <- statswales_get_dataset(
  id,
  filter = list(
    list(Year = c("2022", "2023")),        # AND
    list(Area = c("W92000004"))             # AND
  )
)

Display options

By default the package returns human-readable column names and values. Override with the options argument:

df_raw <- statswales_get_dataset(
  id,
  options = list(
    use_raw_column_names = TRUE,   # internal fact-table column names
    use_reference_values = TRUE,   # reference codes instead of labels
    data_value_type = "raw"        # raw data values
  )
)

Note: as of July 2026 the StatsWales API ignores use_raw_column_names and use_reference_values — output always uses human-readable column names and values. Only data_value_type currently changes the output.

Pivot tables

filters <- statswales_get_filters(id)
pivot <- statswales_get_pivot(
  id,
  x = filters[[1]]$columnName,
  y = filters[[2]]$columnName
)

Dataset metadata

meta <- statswales_get_metadata(id)
meta$published_revision$metadata  # titles and summaries (both languages)
meta$published_revision$designation  # "official", "accredited", etc.

Reusable query IDs

The API stores query configurations with a deterministic 12-character ID. The same filter inputs always produce the same ID, making results shareable and reproducible.

# Create a stored query and inspect it
fid <- statswales_create_query(id, filter = list(list(Year = c("2023"))))
query_info <- statswales_get_query(id, fid)
query_info$totalLines  # total rows matching the query

Download a file

# Download as CSV
path <- statswales_download_dataset(id, format = "csv")
df <- read.csv(path)

# Download filtered data as Excel
statswales_download_dataset(
  id,
  format = "xlsx",
  filter = list(list(Year = c("2022", "2023"))),
  path = "my_data.xlsx"
)

Function reference

Function	Description
`statswales_list_datasets()`	List all published datasets
`statswales_search()`	Full-text search across dataset titles and summaries
`statswales_list_topics()`	List top-level topic categories
`statswales_get_topic()`	Get sub-topics or datasets within a topic
`statswales_get_metadata()`	Full metadata for a dataset
`statswales_get_filters()`	Available filter dimensions and values
`statswales_get_dataset()`	Dataset data as a data frame
`statswales_download_dataset()`	Download dataset as CSV or Excel
`statswales_get_pivot()`	Cross-tabulated pivot view
`statswales_create_query()`	Create a reusable stored query, return filter ID
`statswales_get_query()`	Inspect a stored query configuration

All functions accept a lang parameter: "en-gb" (default), "en", "cy-gb", or "cy".

Hex sticker by Lew Furber.