Get Started with tidyBdE

tidyBdE is an API package that helps to retrieve data from Banco de España. The data is provided as tibble and the package tries to guess the format of every time-series (dates, characters and numbers).

Search series

Banco de España (BdE) provides several time-series, either produced by the institution itself or compiled for another sources, as Eurostat or INE.

The basic entry point for searching time-series are the catalogs (indexes) of information. You can search any series by name:


library(tidyBdE)

# Load tidyverse for better handling
library(tidyverse)


# Search GBP on "TC" (exchange rate) catalog
XR_GBP <- bde_catalog_search("GBP", catalog = "TC")

XR_GBP %>%
  select(Numero_secuencial, Descripcion_de_la_serie) %>%
  # To table on document
  knitr::kable()
Numero_secuencial Descripcion_de_la_serie
573214 Tipo de cambio. Libras esterlinas por euro (GBP/EUR).Datos diarios

Note that BdE files are only provided in Spanish, for the time being, the organism is working on the English version. By now, search terms should be provided in Spanish in order to get search results.

After we have found our series, we can load the series for the GBP/EUR exchange rate using the sequential number reference (Numero_Secuencial) as:


seq_number <- XR_GBP %>%
  # First record
  slice(1) %>%
  # Get id
  select(Numero_secuencial) %>%
  # Convert to num
  as.double()


seq_number
#> [1] 573214

# We can check the metadata of the series
seq_number %>%
  bde_series_load(extract_metadata = TRUE) %>%
  # To nice table for the vignette
  knitr::kable()
Date 573214
NOMBRE DE LA SERIE DTCCBCEGBPEUR.B
NÚMERO SECUENCIAL 573214
ALIAS DE LA SERIE TC_1_1.4
DESCRIPCIÓN DE LA SERIE Tipo de cambio. Libras esterlinas por euro (GBP/EUR).Datos diarios
DESCRIPCIÓN DE LAS UNIDADES Libras esterlinas por Euro
FRECUENCIA LABORABLE


time_series <- bde_series_load(seq_number, series_label = "EUR_GBP_XR") %>%
  filter(Date >= "2010-01-01" & Date <= "2020-12-31") %>%
  drop_na()

Plot series

The package also provides a custom ggplot2 theme based on the publications of BdE:


ggplot(time_series, aes(x = Date, y = EUR_GBP_XR)) +
  geom_line(colour = bde_vivid_pal()(1)) +
  geom_smooth(method = "gam", colour = bde_vivid_pal()(2)[2]) +
  labs(
    title = "EUR/GBP Exchange Rate (2010-2020)",
    subtitle = "%",
    caption = "Source: BdE"
  ) +
  geom_vline(
    xintercept = as.Date("2016-06-23"),
    linetype = "dotted"
  ) +
  geom_label(aes(
    x = as.Date("2016-06-23"),
    y = .95,
    label = "Brexit"
  )) +
  coord_cartesian(ylim = c(0.7, 1)) +
  theme_bde()

plot of chunk chart

The package provides also several “shortcut” functions for a selection of the most relevant macroeconomic series, so there is no need to look for them in advance:


gdp <- bde_ind_gdp_var("values")
gdp$label <- "GDP YoY"

UnempRate <- bde_ind_unemployment_rate("values")
UnempRate$label <- "Unemployment Rate"

plotseries <- bind_rows(gdp, UnempRate) %>%
  drop_na() %>%
  filter(Date >= "2010-01-01" & Date <= "2019-12-31")

ggplot(plotseries, aes(x = Date, y = values)) +
  geom_line(aes(color = label)) +
  labs(
    title = "Spanish Economic Indicators (2010-2019)",
    subtitle = "%",
    caption = "Source: BdE"
  ) +
  theme_bde() +
  scale_color_bde_d() # Custom palette on the package

plot of chunk macroseries

Custom palettes

Two custom palettes, based on the used by BdE on some publications are available:

# Load GDP Series

GDP <- bde_series_load(
  series_code = c(
    3777251,
    3777265,
    3777259,
    3777269,
    3777060
  ),
  series_label = c(
    "Agriculture",
    "Industry",
    "Construction",
    "Services",
    "Total"
  )
)


# Manipulate data - tidyverse style

GDP_all <- GDP %>%
  # Filter dates
  filter(Date <= "2020-12-31") %>%
  # Create 'Other' column and convert Date to year
  mutate(
    Other = Total - rowSums(across(Agriculture:Services)),
    Date = as.numeric(format(Date, format = "%Y"))
  ) %>%
  # Sum by year
  group_by(Date) %>%
  summarise_at(vars(-group_cols()), sum) %>%
  # Create percentage
  relocate(Total, .after = Other) %>%
  mutate(across(Agriculture:Other, ~ .x * 100 / Total)) %>%
  # Move cols to rows for plotting
  select(-Total) %>%
  pivot_longer(Agriculture:Other,
    names_to = "serie",
    values_to = "value"
  )



ggplot(data = GDP_all, aes(
  x = Date,
  y = value,
  fill = serie
)) +
  geom_bar(
    position = "stack",
    stat = "identity",
    alpha = 0.8
  ) +
  scale_fill_bde_d(palette = "bde_rose_pal") + # Custom palette on the package
  scale_x_continuous(expand = c(0, 0)) +
  scale_y_continuous(expand = c(0, 0)) +
  theme_bde() +
  labs(
    title = "Spain: Gross domestic product by industry",
    subtitle = "%",
    caption = "Source: BdE"
  )

plot of chunk gdp

A note on caching

You can use tidyBdE to create your own local repository at a given local directory passing the following option:

options(bde_cache_dir = "./path/to/location")

When this option is set, tidyBdE would look for the cached file on the bde_cache_dir directory and it will load it, speeding up the process.

It is possible to update the data (i.e. after every monthly or quarterly data release) with the following commands:

bde_catalog_update()

# On most of the functions using the option update_cache = TRUE

bde_series_load("SOME ID", update_cache = TRUE)