Skip to contents

Cascade Data to NUTS2 (Generic / Domain-Agnostic)

Source: R/cascade.R
cascade_to_nuts2.Rd

Cascades data from NUTS0/NUTS1 to NUTS2 level for any set of variables from any thematic domain. Unlike cascade_to_nuts2_and_compute(), this function performs pure cascading without computing domain-specific derived indicators, making it suitable for economy, education, labour, environment, or any other Eurostat domain.

When impute = TRUE (the default), adaptive econometric imputation is applied after cascading to fill temporal gaps within each region's time series. This uses PCHIP interpolation for internal gaps and optionally ETS autoregressive forecasting for future periods (when forecast_to is specified). The best model is selected automatically via AIC.

Usage

cascade_to_nuts2(
  data,
  vars,
  years = NULL,
  nuts2_ref = NULL,
  nuts_year = 2024,
  impute = TRUE,
  forecast_to = NULL
)

Arguments

data

Dataframe with 'geo', 'year', and variable columns. May contain data at mixed NUTS levels (identified by geo code length: 2=NUTS0, 3=NUTS1, 4=NUTS2).

vars

Character vector of variable names to cascade

years

Integer vector of years to include. If NULL, uses all available years.

nuts2_ref

Reference table from get_nuts2_ref(). If NULL, fetched automatically (session-cached for instant repeated access).

nuts_year

Integer year for NUTS classification (default: 2024)

impute

Logical. If TRUE (default), apply adaptive econometric imputation (PCHIP + optional ETS) to fill temporal gaps after cascading.

forecast_to

Integer year. If specified and greater than max(years), extend the series with ETS autoregressive forecasts up to this year. Only used when impute = TRUE.

Value

Dataframe with cascaded values at NUTS2 level. For each variable v:

  • src_v_level: source NUTS level (2 = original NUTS2, 1 = cascaded from NUTS1, 0 = cascaded from NUTS0)

  • imp_v_flag: imputation flag (0 = observed/cascaded, 1 = PCHIP interpolated, 2 = ETS forecasted). Only present when impute = TRUE.

Examples

if (FALSE) { # \dontrun{
# Cascade with adaptive imputation (default)
result <- cascade_to_nuts2(
  all_data,
  vars = c("gdp", "unemployment_rate"),
  years = 2015:2024,
  impute = TRUE
)

# Cascade with imputation + forecasting to 2025
result <- cascade_to_nuts2(
  all_data,
  vars = c("gdp", "unemployment_rate"),
  years = 2015:2024,
  impute = TRUE,
  forecast_to = 2025
)

# Check traceability
table(result$src_gdp_level)   # cascade source
table(result$imp_gdp_flag)    # imputation method

# Cascade without imputation (legacy behaviour)
result <- cascade_to_nuts2(
  all_data,
  vars = c("gdp", "unemployment_rate"),
  years = 2015:2024,
  impute = FALSE
)
} # }