Skip to contents

Methodology, sources and warnings

Source: vignettes/articles/en/methodology.Rmd
methodology.Rmd

Important warning on the use of the data

The shape of the municipalities is approximate. For most historical municipalities that no longer exist there is no cartography of their limits. The polygons have been reconstructed approximately and, therefore:

  • The maps are especially useful to represent municipal data graphically.
  • They must be used with caution when geographic variables matter, such as the area of the municipalities, distances to other points, precise contiguity or the exact location of the borders.
  • They are a useful but imperfect approximation, and this must be kept in mind in any analysis.

The same applies to the higher-level units (judicial districts and electoral districts), which are built by aggregating the municipalities.

1. Building the municipal polygons

  1. Starting point. The earliest available municipal division with official boundaries was taken, from the early 2000s (SIANE, Instituto Geográfico Nacional, distributed through the R package mapSpain).
  2. Reconstruction for each census. From this division, a polygon was generated approximately and manually for each municipality existing at each census (maps of 1900, 1930, 1950 and 1970):
    • When a current municipality results from the merger of several historical municipalities, its territory was divided among the former municipalities.
    • The polygon was always drawn so that the municipal seat (the settlement that usually names the municipality) falls inside it.
    • When the settlements that belonged to the municipality were known, they were also kept inside the polygon.
  3. Municipalities in force. The municipalities of each census are those with household figures in the census series of the Instituto Nacional de Estadística (INE).

2. Consistency between censuses and topology

The hand-drawn maps had small overlaps and gaps between polygons (invisible to the naked eye), and the boundaries of the same municipality did not always match exactly between maps. So that aggregations (provinces, judicial districts, districts) have no gaps, the data-building process does the following:

  1. Cleaning each map. Overlapping areas are split equally among the municipalities involved (each point goes to the municipality with the nearest border) and small gaps are filled by equal expansion of the neighbouring borders.
  2. Partition into minimal units. All the boundaries of the 1900, 1930, 1950 and 1970 maps and of SIANE 2005 are overlaid to obtain a single partition of the territory into minimal units. Each unit is assigned a municipality in each map.
  3. Censuses without a hand-drawn map (1910, 1920, 1940, 1960, 1981, 1991 and 2001). Each unit is assigned from the previous and the next map, the municipalities in force at the census and the municipal changes documented in MINAP (Variaciones de los municipios de España desde 1842), to which the INE code has been added. For censuses after 1970 the 1970 boundaries are kept unless a change is documented.
  4. Manual corrections. Some special problematic cases were solved by hand (files data-raw/correccions_manuals.csv, data-raw/parts_desconnectades.csv and data-raw/manual/territoris_manuals.gpkg).
  5. Simplification. The partition is simplified preserving its topology (each shared boundary is simplified only once, with a 20 m tolerance), and the map of each census is obtained by dissolving the units by municipality. Some units are corrected so that they are also valid in geographic coordinates (moving almost coincident vertices a few metres, always in every unit that shares them).

As all maps come from the same partition, they have no gaps or overlaps and any aggregation is exact.

Non-municipal territories (53xxx)

Communal condominiums or jointly administered territories are administrative units that are not municipalities. They are kept as units of their own in the municipal maps (no_munic = 1). In the aggregations:

  • Provinces: they are assigned to the province (SIANE 2005) that contains the polygon.
  • Judicial districts and electoral districts: they are assigned to the unit of their province with which they share the longest boundary.

Canary Islands

Coordinates are real by default. All functions have the option moveCAN = TRUE, which moves the Canary Islands next to the Peninsula exactly as the mapSpain package does. In these maps it is advisable to draw the box that separates the islands, with mapSpain::esp_get_can_box().

3. Higher-level units

Unit Censuses / elections Source
Provinces 1900-2001 INE municipality code (first 2 digits)
Judicial districts 1900, 1910, 1920, 1930 INE censuses (PJ variable) and secondary sources
Districts for the Cortes (Restoration) 1920 (municipalities of the 1920 census) Congreso de los Diputados
Constituencies (Second Republic) 1931, 1933, 1936 (municipalities of the 1930 census) INE censuses and secondary sources to determine the municipalities that belonged to the constituencies of large cities; the rest belonged to the provincial constituency

Unit codes: 2 province digits + 2 unit digits in alphabetical order within the province (judicial districts have the same code in every census). Melilla was not assigned to any district for the Cortes in 1920.

4. Sources

  • Reference municipal boundaries: SIANE (© Instituto Geográfico Nacional, CC BY 4.0), 2005 division, through the R package mapSpain.
  • Municipal maps of 1900, 1930, 1950 and 1970: own reconstruction from the previous source (see section 1).
  • Census figures (1842-2011): population censuses of the INE.
  • Municipal changes: Variaciones de los municipios de España desde 1842 (Ministerio de Administraciones Públicas, MINAP).
  • Electoral districts for the Cortes in 1920: https://ccuc.csuc.cat/permalink/34CSUC_NETWORK/ian09j/alma991020986739706706
  • Judicial districts and constituencies of the Second Republic: information from the INE censuses and other secondary sources.

5. Licenses and citation

  • R package (code and the bundled censos and unitats tables): MIT license.
  • Zenodo data deposit (geometries, census tables and GeoPackage): Creative Commons Attribution 4.0 International license (CC BY 4.0).

If you use the package and/or its data, please cite them (citation("mapSpainHist")):