Este documento analiza en detalle un dataset de Eurostat, ya elegido a partir de las tablas que genera EUROSTAT_TableDatasetGeneration.Rmd, y comprueba si cumple los requisitos para ser el dataset principal del proyecto personal de AEDV.
Para analizar otro dataset basta con cambiar su código en el chunk de abajo y volver a compilar.
# CÓDIGO DEL DATASET QUE VAMOS A ANALIZAR
dataset.name <- "demo_r_pjanaggr3"Descargamos el dataset con la función get_eurostat y lo
ordenamos por fechas y regiones.
# cargamos la tabla y la ordenamos por fechas y regiones
data <- get_eurostat(dataset.name, time_format = "date") |>
as_tibble() |>
arrange(TIME_PERIOD,geo)Antes de guardar nada en disco, eliminamos las columnas que
no aportan ninguna información: aquellas cuyo valor es
siempre el mismo en todas las filas (por ejemplo, freq
suele valer siempre A, porque el dataset ya viene fijado a
una única frecuencia). Se excluyen de esta comprobación
geo, TIME_PERIOD y values, que
son las tres columnas estructurales del dataset (región, fecha y dato
numérico), no dimensiones susceptibles de ser constantes sin más.
## Columnas con un solo valor (se eliminan):
## - freq = A
## - unit = NRGuardamos el dataset lo primero de todo, nada más
limpiarlo, en formato rds (el formato nativo de
R: conserva los tipos de las columnas y ocupa mucho menos
que un Excel).
filename <- dataset.name
saveRDS(data, file = paste0(filename, ".rds"))
cat("Fichero guardado:", paste0(filename, ".rds"),
sprintf("(%.1f MB)\n", file.info(paste0(filename, ".rds"))$size / 1024^2))## Fichero guardado: demo_r_pjanaggr3.rds (2.4 MB)Guardamos también los metadatos del dataset recogidos en el catálogo
de la asignatura (Eurostat.datasets.xlsx, generado por
EUROSTAT_TableDatasetGeneration.Rmd): la fila
correspondiente a este dataset, con las columnas pasadas a filas
(Campo/Valor) para que quepan en una tabla
estrecha.
# el catálogo habitual es el filtrado; si el dataset no está en él (el filtro
# deja fuera datasets que sí sirven) se busca en el volcado completo .raw
fila_metadatos <- openxlsx::read.xlsx("Eurostat.datasets.xlsx") |>
filter(code == dataset.name)
if(nrow(fila_metadatos) == 0)
fila_metadatos <- openxlsx::read.xlsx("Eurostat.datasets.raw.xlsx") |>
filter(code == dataset.name)
if(nrow(fila_metadatos) == 0)
stop("El dataset '", dataset.name, "' no aparece en el catálogo de Eurostat. ",
"Comprueba el código o vuelve a generar los catálogos con EUROSTAT_TableDatasetGeneration.Rmd")
# el catálogo solo trae `topic` (nombre corto de la familia, para no cargar cada
# fila con miles de caracteres repetidos); aquí recuperamos la descripción
# completa desde Eurostat.topics.xlsx, para que quede en los metadatos del dataset
topics_tabla <- openxlsx::read.xlsx("Eurostat.topics.xlsx") |> as_tibble()
fila_metadatos <- fila_metadatos |>
left_join(topics_tabla, by = "topic")
# pasamos las columnas a filas: son 26 campos, mucho más legibles en vertical
metadatos <- tibble(
Campo = colnames(fila_metadatos),
Valor = as.character(t(fila_metadatos))
)
openxlsx::write.xlsx(metadatos, file = paste0(filename, "_metadatos.xlsx"), overwrite = TRUE)Una vez guardados en disco los metadatos (fichero xlsx) y los datos
(fichero RDS), iniciamos el proceso de análisis. A partir de aquí el
documento se compila solo con los ficheros ya guardados
en disco (el rds y el xlsx de metadatos); no
se usa ningún objeto de las secciones anteriores. Es a propósito: es la
parte que se incorpora a la memoria del proyecto personal (ver la última
sección, Incorporación del dataset a la memoria), así que tiene
que poder ejecutarse sola (lo único que se usa de lo definido
anteriormente es el dataset.name)
metadatos
| Campo | Valor |
|---|---|
| code | demo_r_pjanaggr3 |
| title | Population on 1 January by broad age group, sex and NUTS 3 region |
| topic | POPULATION_DEMOGRAPHY_REGIONAL |
| Freq | Y |
| N.periodos | 36 |
| Ini | 1990 |
| Fin | 2025 |
| N.observaciones | 735652 |
| N.regiones | 2195 |
| NUTS0 | 36 |
| NUTS1 | 113 |
| NUTS2 | 295 |
| NUTS3 | 1338 |
| AGREG | 4 |
| OTHERS | 409 |
| ESPAÑA | TRUE |
| CANARIAS | TRUE |
| N.atrib.combinados | 15 |
| Dimension_N.options | unit 1 - sex 3 - age 5 |
| ESMS | https://ec.europa.eu/eurostat/cache/metadata/en/demo_r_gind3_esms.htm |
| last.update.of.data | 28.08.2026 |
| last.table.structure.change | 10.08.2026 |
| data.start | 1990 |
| data.end | 2025 |
| hierarchy | 5 |
Descripción del topic al que pertenece el dataset
Each year Eurostat collects demographic data at regional level from EU, EFTA and Candidate countries as part of the Population Statistics data collection. POPSTAT is Eurostat’s main annual demographic data collection and aims to gather information on demography and migration at national and regional levels by various breakdowns (for the full overview see the Eurostat dedicated section). More specifically, POPSTAT collects data at regional levels on: population stocks; vital events (live births and deaths). Each country must send the statistics for the reference year (T) to Eurostat by 31 December of the following calendar year (T+1). Eurostat then publishes the data in March of the calendar year after that (T+2). Demographic data at regional level include statistics on the population at the end of the calendar year and on live births and deaths during that year, according to the official classification for statistics at regional level (NUTS - nomenclature of territorial units for statistics) in force in the year. These data are broken down by NUTS 2 and 3 levels for EU countries. For more information on the NUTS classification and its versions please refer to the Eurostat dedicated pages. For EFTA and Candidate countries the data are collected according to the agreed statistical regions that have been coded in a way that resembles NUTS. The breakdown of demographic data collected at regional level varies depending on the NUTS/statistical region level. These breakdowns are summarised below, along with the link to the corresponding online table: NUTS 2 level Population by sex, age and region of residence — demo_r_d2jan Population on 1 January by age group, sex and region of residence — demo_r_pjangroup Live births by mother’s age, mother’s year of birth and mother’s region of residence — demo_r_fagec Deaths by sex, age, and region of residence — demo_r_magec NUTS 3 level Population on 1 January by sex, age group and region of residence — demo_r_pjangrp3 Population on 1 January by broad age group, sex and region of residence — demo_r_pjanaggr3 Live births (total) by region of residence — demo_r_births Live births by five-year age group of the mothers and region of residence — demo_r_fagec3 Deaths (total) by region of residence — demo_r_deaths Deaths by five-year age group, sex and region of residence — demo_r_magec3 This more detailed breakdown (by five-year age group) of the data collected at NUTS 3 level started with the reference year 2013 and is in accordance with the European laws on demographic statistics. In addition to the regional codes set out in the NUTS classification in force, these online tables include few additional codes that are meant to cover data on persons and events that cannot be allocated to any official NUTS region. These codes are denoted as CCX/CCXX/CCXXX (Not regionalised/Unknown level 1/2/3; CC stands for country code) and are available only for France, Hungary, North Macedonia and Albania, reflecting the raw data as transmitted to Eurostat. For the reference years from 1990 to 2012 all countries sent to Eurostat all the data on a voluntary basis, therefore the completeness of the tables and the length of time series reflect the level of data received from the responsible National Statistical Institutes’ (NSIs) data provider. As a general remark, a lower data breakdown is available at NUTS 3 level as detailed: population data are broken down by sex and broad age groups (0-14, 15-64 and 65 or more). The data have this disaggregation since the reference year 2007 for all countries, and even longer for some — demo_r_pjanaggr3 vital events (live births and deaths) data are available only as totals, without any further breakdown — demo_r_births and demo_r_deaths Demographic indicators are calculated by Eurostat based on the above raw data using a common methodology for all countries and regions. The regional demographic indicators computed by NUTS level and the corresponding online tables are summarised below: NUTS 2 level Population structure indicators by region of residence (shares of various population age groups, dependency ratios and median age) — demo_r_pjanind2 Fertility indicators by region of residence — demo_r_find2 Fertility rates by age and region of residence — demo_r_frate2 Life table by age, sex and region of residence — demo_r_mlife Life expectancy by age, sex and region of residence — demo_r_mlifexp Infant mortality rates by region of residence — demo_r_minfind NUTS 3 level Population change - Demographic balance and crude rates at regional level — demo_r_gind3 Population density by region — demo_r_d3dens Population structure indicators by region of residence (shares of various population age groups, dependency ratios and median age) — demo_r_pjanind3 Fertility indicators by region of residence (total fertility rate, mean age of woman at childbirth and median age of woman at childbirth) — demo_r_find3 Notes: 1) All the indicators are computed for all lower NUTS regions included in the tables (e.g. data included in a table at NUTS 3 level will include also the data for NUTS 2, 1 and country levels). 2) Demographic indicators computed by NUTS 2 and 3 levels are calculated using input data that have different age breakdown. Therefore, minor differences can be noted between the values corresponding to the same indicator of the same region classified as NUTS 2, 1 or country level. 3) Since the reference year 2015, Eurostat has stopped collecting data on area; therefore, the table ‘Area by NUTS 3 region (demo_r_d3area)’ includes data up to the year 2015 included. 4) Starting with the reference year 2016, the population density indicator is computed using the new data on area ’Area by NUTS 3 region (reg_area3).
data
## # A tibble: 735,709 × 5
## sex age geo TIME_PERIOD values
## <chr> <chr> <chr> <date> <dbl>
## 1 T TOTAL AL 1990-01-01 3286500
## 2 T TOTAL AL0 1990-01-01 3286500
## 3 F TOTAL AT 1990-01-01 3989903
## 4 F UNK AT 1990-01-01 0
## 5 F Y15-64 AT 1990-01-01 2589567
## # ℹ 735,704 more rowsSe presentan las diferentes dimensiones, cada una con sus valores posibles y el nº de observaciones.
COLUMNA: sex (3 opciones)
sex N.Observ full_name
[1,] T 245247 Total
[2,] F 245231 Females
[3,] M 245231 Males COLUMNA: age (5 opciones)
age N.Observ full_name
[1,] TOTAL 165915 Total
[2,] UNK 145060 Unknown
[3,] Y_GE65 141594 65 years or over
[4,] Y15-64 141570 From 15 to 64 years
[5,] Y_LT15 141570 Less than 15 years CONTABILIZACIÓN Nº REGIONES NUTS A PARTIR DE LA COLUMNA geo (2195 opciones)
NUTS Number of Regions
[1,] 0 36
[2,] 1 113
[3,] 2 295
[4,] 3 1338
[5,] EFTA 1
[6,] EU27_2007 1
[7,] EU27_2020 1
[8,] EU28 1
[9,] OTHERS 409 COLUMNA: TIME_PERIOD (36 opciones)
TIME_PERIOD N.observ
[1,] 1990-01-01 4485
[2,] 1991-01-01 5432
[3,] 1992-01-01 6092
[4,] 1993-01-01 5354
[5,] 1994-01-01 5290
[6,] 1995-01-01 5594
[7,] 1996-01-01 5380
[8,] 1997-01-01 5383
[9,] 1998-01-01 5384
[10,] 1999-01-01 5471
[11,] 2000-01-01 13117
[12,] 2001-01-01 16443
[13,] 2002-01-01 19845
[14,] 2003-01-01 20994
[15,] 2004-01-01 20973
[16,] 2005-01-01 21030
[17,] 2006-01-01 25497
[18,] 2007-01-01 26391
[19,] 2008-01-01 27885
[20,] 2009-01-01 27915
[21,] 2010-01-01 28413
[22,] 2011-01-01 28725
[23,] 2012-01-01 29463
[24,] 2013-01-01 23232
[25,] 2014-01-01 30688
[26,] 2015-01-01 31623
[27,] 2016-01-01 31623
[28,] 2017-01-01 31500
[29,] 2018-01-01 31542
[30,] 2019-01-01 31464
[31,] 2020-01-01 27986
[32,] 2021-01-01 27426
[33,] 2022-01-01 27615
[34,] 2023-01-01 27234
[35,] 2024-01-01 26460
[36,] 2025-01-01 26760 Nº de observaciones en España/Canarias
## # A tibble: 10 × 3
## geo full_name N.Observ
## <chr> <chr> <int>
## 1 ES Spain 540
## 2 ES7 Canarias 540
## 3 ES70 Canarias 540
## 4 ES703 El Hierro 360
## 5 ES704 Fuerteventura 360
## 6 ES705 Gran Canaria 360
## 7 ES706 La Gomera 360
## 8 ES707 La Palma 360
## 9 ES708 Lanzarote 360
## 10 ES709 Tenerife 360Combinaciones de atributos: cobertura por nivel NUTS y percentiles del tamaño de la serie temporal (nº de periodos por región).
| sex | age | n_regiones | NUTS0 | NUTS1 | NUTS2 | NUTS3 | P0 | P25 | P50 | P75 | P100 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| F | TOTAL | 2195 | 36 | 113 | 295 | 1338 | 1 | 23 | 25 | 31 | 36 |
| M | TOTAL | 2195 | 36 | 113 | 295 | 1338 | 1 | 23 | 25 | 31 | 36 |
| T | TOTAL | 2195 | 36 | 113 | 295 | 1338 | 1 | 23 | 25 | 31 | 36 |
| F | UNK | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 22 | 26 | 36 |
| F | Y15-64 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| F | Y_GE65 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| F | Y_LT15 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| M | UNK | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 22 | 26 | 36 |
| M | Y15-64 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| M | Y_GE65 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| M | Y_LT15 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| T | UNK | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 22 | 26 | 36 |
| T | Y15-64 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| T | Y_GE65 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| T | Y_LT15 | 2195 | 36 | 113 | 295 | 1338 | 1 | 18 | 19 | 26 | 36 |
| Atributos | Cobertura geográfica | Datos actualizados (≥ 2024) | Tamaño series (P25 ≥ umbral) | Tamaño series España/Canarias | CUMPLE LOS 4 |
|---|---|---|---|---|---|
| F / TOTAL | ✔ | ✔ | ✔ | ✔ | ✔ |
| M / TOTAL | ✔ | ✔ | ✔ | ✔ | ✔ |
| T / TOTAL | ✔ | ✔ | ✔ | ✔ | ✔ |
| F / UNK | ✔ | ✔ | ✔ | ✔ | ✔ |
| F / Y15-64 | ✔ | ✔ | ✔ | ✔ | ✔ |
| F / Y_GE65 | ✔ | ✔ | ✔ | ✔ | ✔ |
| F / Y_LT15 | ✔ | ✔ | ✔ | ✔ | ✔ |
| M / UNK | ✔ | ✔ | ✔ | ✔ | ✔ |
| M / Y15-64 | ✔ | ✔ | ✔ | ✔ | ✔ |
| M / Y_GE65 | ✔ | ✔ | ✔ | ✔ | ✔ |
| M / Y_LT15 | ✔ | ✔ | ✔ | ✔ | ✔ |
| T / UNK | ✔ | ✔ | ✔ | ✔ | ✔ |
| T / Y15-64 | ✔ | ✔ | ✔ | ✔ | ✔ |
| T / Y_GE65 | ✔ | ✔ | ✔ | ✔ | ✔ |
| T / Y_LT15 | ✔ | ✔ | ✔ | ✔ | ✔ |
De las 15 combinaciones de atributos analizadas, 15 cumplen los cuatro criterios de validación (cobertura geográfica, datos actualizados, tamaño de las series temporales y tamaño de las series temporales en España/Canarias).
El requisito de disponer de al menos 4 atributos válidos se cumple ✓.
Al compilar este documento se ha generado el fichero
demo_r_pjanaggr3_analisis.Rmd: es la
sección Análisis del dataset de arriba, preparada para la
sección Comprensión de los datos de la memoria, con los
títulos al nivel de esa sección. Para incorporar el dataset a la
memoria:
EUROSTAT: demo_r_pjanaggr3.rds,
demo_r_pjanaggr3_metadatos.xlsx y
demo_r_pjanaggr3_analisis.Rmd.utilidades.R (viene en la carpeta
ProyectoPersonal, junto al modelo de memoria) y que la
memoria lo carga en el chunk de librerías:
source("utilidades.R").```{r, results='asis', cache=FALSE}
datos_demo_r_pjanaggr3 <- aedv_incluir_dataset("demo_r_pjanaggr3_analisis.Rmd")
```Al compilar la memoria, el análisis aparece en el lugar del chunk y
los datos quedan en la tabla datos_demo_r_pjanaggr3 (un
tibble), con la que se sigue trabajando en el resto de la memoria. Si
incorporas varios datasets, pega un chunk por cada uno, con nombres de
tabla distintos. En un dataset auxiliar, que no tiene que cumplir los
requisitos de AEDV,
aedv_incluir_dataset("demo_r_pjanaggr3_analisis.Rmd", validacion = FALSE)
omite la validación.
El texto explicativo del dataset (qué mide, qué significa cada
variable y cada una de sus opciones…) se escribe en la memoria, antes o
después del chunk, no en demo_r_pjanaggr3_analisis.Rmd: ese
fichero se sustituye cada vez que se vuelve a generar.