pib_ccaa.R

# Dades per CCAA ===============================================================

library(tidyverse)
library(rvest)
theme_set(theme_minimal())
older_year <- 2002

region <- tibble(
    slug = c("andalucia", "aragon", "asturias", "canarias", "cantabria", "castilla-leon", 
                 "castilla-la-mancha", "cataluna", "ceuta", "melilla", "madrid", "valencia", 
                 "extremadura", "galicia", "islas-baleares", "la-rioja", "navarra", "murcia", 
                 "pais-vasco"),
    es = c("Andalucía", "Aragón", "Asturias", "Canarias", "Cantabria", "Castilla y León", 
                 "Castilla-La Mancha", "Cataluña", "Ceuta", "Melilla", "Madrid", "Comunidad Valenciana", 
                 "Extremadura", "Galicia", "Islas Baleares", "La Rioja", "Navarra", "Región de Murcia", 
                 "País Vasco"),
    name = c("Andalusia", "Aragó", "Astúries", "Canàries", "Cantàbria", "Castella i Lleó", 
             "Castella-la Manxa", "Catalunya", "Ceuta", "Melilla", "Madrid", "Comunitat Valenciana", 
             "Extremadura", "Galícia", "Illes Balears", "La Rioja", "Navarra", "Regió de Múrcia", 
             "País Basc")
)
safe_html <- safely(read_html, otherwise = NULL)
if(!exists("html_results")) {
    html_results <- 
        region$slug %>%
        map(\(x) {
            tibble(
                slug = x,
                tables = safe_html(glue::glue("https://datosmacro.expansion.com/pib/espana-comunidades-autonomas/{x}")) %>% 
                    pluck("result") %>% 
                    html_table()
            )
        })
}

tables <- 
    html_results %>%
    list_rbind() %>% 
    mutate(tables = map(tables, \(x)janitor::clean_names(x))) %>% 
    mutate(row = paste0("table", 1 + (1 +row_number()) %% 2))

money_to_dobule <- \(x){
    x %>%
        str_replace("M", "000000") %>%
        str_replace_all("(\\s| |\\.|€)", "") %>%
        parse_double()
}


## pib_by_ccaa =================================================================

pib_by_ccaa <- 
    tables %>% 
    filter(str_detect(row, 'table1')) %>%
    pivot_wider(id_cols = slug, names_from = row, values_from = tables) %>%
    unnest(table1) %>%
    transmute(
        slug,
        year_integer = fecha,
        year_date = make_date(year = fecha),
        pib_anual = money_to_dobule(pib_anual),
    ) %>% 
    filter(year_integer >= older_year) %>% 
    left_join(region, by = join_by(slug))

pib_by_ccaa_labels <- filter(pib_by_ccaa, near(year_integer, 2022), pib_anual > 50000000000)

pib_by_ccaa %>% 
    ggplot(aes(year_date, pib_anual, colour = name)) +
    geom_line() +
    ggrepel::geom_text_repel(aes(label = name), data = pib_by_ccaa_labels)  +
    scale_x_date(date_breaks = "year", date_labels ="%Y") +
    scale_y_continuous(
        labels = scales::unit_format(unit = "M€", scale = 1e-6)
    ) +
    labs(
        title = "Evolució anual del PIB anual per CCAA", 
        x = "", y = "", colour = "CCAA",
        caption = "Autor @jrosell | Font https://datosmacro.expansion.com"
    ) +
    theme(legend.position = "none")


## pib_per_capita_by_ccaa ======================================================

pib_per_capita_by_ccaa <- 
    tables %>% 
    filter(str_detect(row, 'table2')) %>%
    pivot_wider(id_cols = slug, names_from = row, values_from = tables) %>%
    unnest(table2) %>%
    transmute(
        slug,
        year_integer = fecha,
        year_date = make_date(year = fecha),
        pib_per_capita = money_to_dobule(pib_per_capita),
    ) %>% 
    filter(year_integer >= older_year) %>% 
    left_join(region, by = join_by(slug)) %>% 
    select(-slug, -es)

pib_per_capita_by_ccaa %>% head()

pib_per_capita_by_ccaa_labels_first <- 
    pib_per_capita_by_ccaa %>% 
    filter(near(year_integer, 2002) & pib_per_capita > 19000)

pib_per_capita_by_ccaa_labels_last <- 
    pib_per_capita_by_ccaa %>% 
    filter(near(year_integer, 2022) & pib_per_capita > 28000)

pib_per_capita_by_ccaa %>% 
    ggplot(aes(year_date, pib_per_capita, colour = name)) +
    geom_line() +
    ggrepel::geom_text_repel(aes(label = name), data = pib_per_capita_by_ccaa_labels_first, nudge_x = -400) +
    ggrepel::geom_text_repel(aes(label = name), data = pib_per_capita_by_ccaa_labels_last, nudge_x = 400)  +
    scale_x_date(
        breaks = seq(as.Date("2002-01-01"), as.Date("2022-12-31"), by = "years"),
        date_labels ="%Y"
    ) +
    scale_y_continuous(
        limits = c(10000, 40000),
        breaks = seq(10000, 40000, by = 5000),
        labels = scales::label_number(big.mark = ".", decimal.mark = ",", suffix = "€")
    ) +
    labs(
        title = "Evolució anual del PIB per càpita per CCAA", 
        x = "", y = "", colour = "CCAA",
        caption = "Autor @jrosell | Font https://datosmacro.expansion.com"
    ) +
    theme(
        legend.position = "none",
        panel.grid.major.y = element_line(color = "gainsboro"),
        panel.grid.minor.y = element_blank(),
        panel.grid.major.x = element_line(color = "gainsboro"),
        panel.grid.minor.x = element_blank(),
    )