Skip to contents

Getting Started with localintel

Source: vignettes/localintel.Rmd
localintel.Rmd

Overview

localintel (Local Intelligence) is an R package that provides a unified pipeline for 150+ subnational indicators across 14 thematic domains from Eurostat. Any regional dataset — economy, health, education, labour, demographics, tourism, transport, environment, and more — can be fetched, harmonized, cascaded to NUTS 2, scored, mapped, and exported through a single consistent workflow.

Installation 

# install.packages("devtools")
devtools::install_github("MohamedHtitich1/localintel")

Step 1: Browse the Indicator Registry

localintel ships with curated code lists for 14 Eurostat domains:

# How many indicators are available?
n <- indicator_count()
cat(n$indicators, "indicators across", n$domains, "domains\n")
print(n$by_domain)

# View codes for a specific domain
economy_codes()
labour_codes()
education_codes()

Step 2: Fetch Data from Eurostat

Use fetch_eurostat_batch() to download multiple datasets at once:

# Fetch economy indicators at NUTS2 level
econ_data <- fetch_eurostat_batch(economy_codes(), level = 2, years = 2015:2024)
econ_data <- drop_empty(econ_data)

# Fetch health indicators
hlth_data <- fetch_eurostat_batch(health_system_codes(), level = 2, years = 2015:2024)
hlth_data <- drop_empty(hlth_data)

# Fetch labour indicators
lab_data <- fetch_eurostat_batch(labour_codes(), level = 2, years = 2015:2024)
lab_data <- drop_empty(lab_data)

For individual datasets with more control:

gdp_raw <- get_nuts_level_robust("nama_10r_2gdp", level = 2, years = 2015:2024)

Step 3: Process Datasets

Use domain-specific processors or the universal process_eurostat():

# Domain-specific processors
gdp       <- process_gdp(econ_data$gdp_nuts2)
beds      <- process_beds(hlth_data$beds)
unemp     <- process_unemployment_rate(lab_data$unemployment_rate)

# Or the generic processor for any dataset
life_exp <- process_eurostat(
  get_nuts_level_robust("demo_r_mlifexp", level = 2, years = 2015:2024),
  filters = list(sex = "T", age = "Y_LT1"),
  out_col = "life_expectancy"
)

Combine processed datasets:

all_data <- merge_datasets(gdp, beds, unemp, life_exp)

Step 4: Cascade Data to NUTS 2

Not every region reports at NUTS 2. The cascade fills gaps from parent levels:

nuts2_ref <- get_nuts2_ref()

cascaded <- cascade_to_nuts2(
  all_data,
  vars = c("gdp", "beds", "unemployment_rate", "life_expectancy"),
  nuts2_ref = nuts2_ref,
  years = 2015:2024
)

# Check source levels — how much data was cascaded?
table(cascaded$src_gdp_level)

Step 5: Visualize 

geopolys <- get_nuts_geopolys()

plot_best_by_country_level(
  cascaded, geopolys,
  var = "unemployment_rate",
  years = 2022:2024,
  title = "Unemployment Rate (%)",
  scale = "global"
)

Step 6: Export 

# Single variable
sf_data <- build_display_sf(cascaded, geopolys, var = "gdp", years = 2015:2024)
export_to_geojson(sf_data, "output/gdp_nuts2.geojson")

# Multi-variable with labels and domain grouping
sf_all <- build_multi_var_sf(
  cascaded, geopolys,
  vars = c("gdp", "beds", "unemployment_rate", "life_expectancy"),
  years = 2015:2024,
  var_labels = regional_var_labels(),
  pillar_mapping = regional_domain_mapping()
)

sf_enriched <- enrich_for_tableau(sf_all, get_population_nuts2(), get_nuts2_names())
export_to_geojson(sf_enriched, "output/multi_domain_dashboard.geojson")

Data Sources

localintel works with the full Eurostat regional statistics catalogue. Use all_regional_codes() for the complete list, or domain-specific functions like economy_codes(), education_codes(), tourism_codes(), etc.