tidybrregpanel

Managed state and longitudinal analysis for Norway's business registry, built on tidybrreg.

tidybrreg is the typed client: sixteen functions, one per register endpoint, no persistence. tidybrregpanel is everything that keeps state on disk or derives quantities over time:

Area	Functions
Snapshot store (Hive-partitioned Parquet)	brreg_snapshot(), brreg_import(), brreg_snapshots(), brreg_open(), brreg_cleanup(), brreg_data_dir(), brreg_manifest(), brreg_status()
Synchronised mirror with changelog	brreg_sync(), brreg_sync_status(), brreg_changes(), brreg_change_summary()
Reconstruction and events	brreg_replay(), brreg_events(), diff_roller_state()
Panels and series	brreg_panel(), brreg_flows(), brreg_series(), as_brreg_tsibble()
Annotations and name history	brreg_annotations(), brreg_annotation_summary(), brreg_historical_names()
Harmonisation (SSB KLASS)	brreg_harmonize_kommune(), brreg_harmonize_nace()
Governance	brreg_board_summary(), brreg_board_network(), brreg_network(), brreg_survival_data()

All retrieval goes through tidybrreg, and internal parsers are reached through tidybrreg::: (the one deliberate ::: dependency in the ecosystem, so the core keeps its exact sixteen exports).

Installation

install.packages(c("tidybrreg", "tidybrregpanel"),
  repos = c("https://sondreskarsten.r-universe.dev", "https://cloud.r-project.org"))

# A parquet backend
install.packages("nanoparquet")  # lightweight
install.packages("arrow")        # lazy queries via brreg_open()

Quick start

library(tidybrreg)
library(tidybrregpanel)

brreg_snapshot("enheter")
brreg_snapshot("enheter")                    # JSON bulk: kapital.*, historiskeNavn, annotations
brreg_snapshots()

panel <- brreg_panel(frequency = "year", cols = c("employees", "nace_1", "legal_form"))
brreg_series(.vars = "employees", by = "legal_form") |> as_brreg_tsibble()

brreg_sync(types = c("enheter", "roller"))
brreg_changes(registry = "roller", change_type = "entry")

See the vignettes: panels, roller-cdc, governance, architecture, and ARCHITECTURE.md for the data flow from the API through parsing, storage, panel construction and time series.

Status

This package carries the managed subsystem exactly as it was in tidybrreg 0.5.0 (commit c9bd992), moved rather than rewritten. The defects recorded against that subsystem in the stress-test evaluation (flows bias D-50..D-52, sync loss D-36/D-38/D-43, harmonisation D-78..D-80, snapshot layout D-54, manifest D-55/D-58/D-102) are open work here, tracked in the tidybrreg issue tracker with the scope:deferred label. The package is not CRAN-bound.