R/etl_labor.R

In schoulten/macroview: Dashboard: Macroeconomic Data of Brazil

#' ETL Labor market
#'
#' @encoding UTF-8
#' @import dplyr PNADcIBGE
#' @importFrom PNADcIBGE pnadc_labeller
#' @importFrom utils lsf.str
#' @return RDATA
#' @export
#'
etl_labor <- function(){

### Labor market ###


# This R code provides the extraction and data wrangling of macroeconomic
# variables referring to the Brazilian labour market


# Packages ----------------------------------------------------------------


# Install/load packages
# if (!require("pacman")) install.packages("pacman")
# pacman::p_load(
#   "PNADcIBGE",
#   "tidyverse",
#   "lubridate",
#   "sidrar",
#   "janitor",
#   "survey",
#   "convey",
#   "rvest",
#   "here"
#   )

if (!"package:PNADcIBGE" %in% search()) { attachNamespace("PNADcIBGE") }


# Parameters --------------------------------------------------------------


# Parameters used in the code to import, download or cleaning data


# Load useful functions
# source("./R/utils.R")


# List of parameters to get data from IBGE website
api_ibge <- dplyr::lst(

  # Continuous National Household Sample Survey - Quarterly Disclosure (PNADC-T)

  # Query variables
  api_pnadc_variables = c(
    "Ano", "Trimestre", "UF", "V1028", "V2007", "V2009", "V2010",
    "V3007", "VD4001", "VD4002", "VD4009", "VD4020", "VD4035"
    ),

  # Last available quarter and year
  api_pnadc_dates = xml2::read_html("https://sidra.ibge.gov.br/home/pnadct/brasil") %>%
    rvest::html_nodes(".titulo-aba") %>%
    rvest::html_text() %>%
    stringr::str_extract(" - \\d{1}. trimestre \\d{4}") %>%
    dplyr::lst(
      quarter = stringr::str_extract(., "\\d{1}") %>% as.numeric(),
      year = stringr::str_extract(., "\\d{4}") %>% as.numeric()
      ),


  # Unemployment rate (PNADC-T)
  api_unemployment = "/t/4093/n1/all/v/4090,4092,4099/p/all/c2/6794/d/v4099%201",

  # Employment by occupation / category in the main job (PNADC-T)
  api_employment = "/t/4097/n1/all/v/4090/p/all/c11913/31721,31724,31727,31731,96170,96171",

  # Labor force participation rate and occupation level (PNADC-T)
  api_labour_force = "/t/4093/n1/all/v/4096,4097/p/all/c2/all/d/v4096%201,v4097%201",

  # Average real income from work, by age group (PNADC-T)
  api_income = "/t/5437/n1/all/v/5935/p/all/c58/all"

  )



# Import data -------------------------------------------------------------


# This section performs the import of data from different sources


# PNADC-T microdata from IBGE
raw_pnadct <- PNADcIBGE::get_pnadc(
  year    = api_ibge$api_pnadc_dates$year,
  quarter = api_ibge$api_pnadc_dates$quarter,
  vars    = api_ibge$api_pnadc_variables,
  design  = FALSE
  )

# Unemployment rate (PNADC-T)
raw_unemployment <- sidrar::get_sidra(api = api_ibge$api_unemployment)

# Employment by occupation / category in the main job (PNADC-T)
raw_employment <- sidrar::get_sidra(api = api_ibge$api_employment)

# Labor force participation rate and occupation level (PNADC-T)
raw_labor_force <- sidrar::get_sidra(api = api_ibge$api_labour_force)

# Average real income from work, by age group (PNADC-T)
raw_income <- sidrar::get_sidra(api = api_ibge$api_income)




# Data wrangling ----------------------------------------------------------


# This section performs data wrangling


# Number of employed and unemployed people
big_numbers <- raw_pnadct %>%
  group_by(VD4002) %>%
  summarise(n = round(sum(V1028) / 1e6, 1)) %>%
  tidyr::drop_na() %>%
  pivot_wider(names_from = VD4002, values_from = n) %>%
  rename_with(~c("employed", "unemployed")) %>%
  mutate(
    across(everything(), ~paste0(., " million")),
    date = paste0(api_ibge[[2]][[3]], " Q", api_ibge[[2]][[2]])
    )


# Employment by category in the main job
employment_category <- raw_employment %>%
  select(
    date     = "Trimestre (C\u00f3digo)",
    variable = "Posi\u00e7\u00e3o na ocupa\u00e7\u00e3o e categoria do emprego no trabalho principal",
    value    = "Valor"
    ) %>%
  filter(date == max(as.numeric(date))) %>%
  mutate(
    date     = stringr::str_replace(date, "(\\d{4})0(\\d{1}$)", "\\1 Q\\2"),
    variable = recode(
      variable,
      "Empregado no setor privado, exclusive trabalhador dom\u00e9stico" = "Employee in the private sector",
      "Trabalhador dom\u00e9stico"         = "Domestic worker",
      "Empregado no setor p\u00fablico"    = "Employee in the public sector",
      "Trabalhador familiar auxiliar" = "Auxiliary family worker",
      "Empregador"                    = "Employer",
      "Conta pr\u00f3pria"                 = "Own account",
      ),
    value    = (value / 1e3) %>% round(2)
    ) %>%
  arrange(desc(value))
#num_txt_ocupacao <- c("Mil", "Milhão", "Milhões", "Bilhão", "Bilhões")


# Unemployment rate
unemployment <- raw_unemployment %>%
  select(
    date     = "Trimestre (C\u00f3digo)",
    variable = "Vari\u00e1vel",
    value    = "Valor"
    ) %>%
  mutate(
    date     = stringr::str_replace(date, "(\\d{4})0(\\d{1}$)", "\\1 Q\\2"),
    region   = "Brazil",
    variable = recode(
      variable,
      "Pessoas de 14 anos ou mais de idade, ocupadas na semana de refer\u00eancia" = "employed",
      "Pessoas de 14 anos ou mais de idade, desocupadas na semana de refer\u00eancia" = "unemployed",
      "Taxa de desocupa\u00e7\u00e3o, na semana de refer\u00eancia, das pessoas de 14 anos ou mais de idade" = "unemployment_rate"
      )
    ) %>%
  pivot_wider(
    id_cols     = c(date, region),
    names_from  = variable,
    values_from = value
    ) %>%
  slice_tail(n = 24) %>%
  mutate(across(employed:unemployed, ~ . * 1e3))


# Unemployment rate by gender and color
unemployment_gender_color <- raw_pnadct %>%
  group_by(
    "gender" = V2007,
    "color" = V2010,
    VD4002
    ) %>%
  summarise(n = sum(V1028)) %>%
  tidyr::drop_na() %>%
  ungroup() %>%
  pivot_wider(
    names_from  = VD4002,
    values_from = n
    ) %>%
  rename_with(~c("employed", "unemployed"), 3:4) %>%
  tidyr::replace_na(list(unemployed = 0)) %>%
  group_by(gender, color) %>%
  mutate(
    unemployment_rate = (unemployed / (unemployed + employed) * 100) %>% round(2),
    date              = paste0(api_ibge[[2]][[3]], " Q", api_ibge[[2]][[2]]),
    gender = recode(gender, "Homem" = "Man", "Mulher" = "Woman"),
    color = recode(
      color,
      "Branca" = "White",
      "Preta" = "Black",
      "Amarela" = "Yellow",
      "Parda" = "Brown",
      "Ind\u00edgena" = "Indian"
      )
    ) %>%
  filter(color != "Ignorado")


# Create attributes for the plot of the last object
gender_attr <- data.frame(
  gender = c("Man", "Woman"),
  color  = c("#2980b9", "#9b59b6"),
  icon   = c(
    here::here("inst/imgs/man.png"),
    here::here("inst/imgs/woman.png")
    )
  )


# Create list of Brazilian states for the map plot
states <- tibble(
  states = c(
    "Rond\u00f4nia", "Acre", "Amazonas", "Roraima", "Par\u00e1", "Amap\u00e1",
    "Tocantins", "Maranh\u00e3o", "Piau\u00ed", "Cear\u00e1", "Rio Grande do Norte",
    "Para\u00edba", "Pernambuco", "Alagoas", "Sergipe", "Bahia",
    "Minas Gerais", "Esp\u00edrito Santo", "Rio de Janeiro", "S\u00e3o Paulo",
    "Paran\u00e1", "Santa Catarina", "Rio Grande do Sul",
    "Mato Grosso do Sul", "Mato Grosso", "Goi\u00e1s", "Distrito Federal"
    ),
  code = c(
    "RO", "AC", "AM", "RR", "PA", "AP", "TO", "MA", "PI",
    "CE", "RN", "PB", "PE", "AL", "SE", "BA", "MG", "ES",
    "RJ", "SP", "PR", "SC", "RS", "MS", "MT", "GO", "DF"
    )
  )


# Unemployment rate by states
unemployment_states <- raw_pnadct %>%
  group_by(
    "states" = UF,
    VD4002
    ) %>%
  summarise(n = sum(V1028)) %>%
  tidyr::drop_na() %>%
  ungroup() %>%
  pivot_wider(
    names_from  = VD4002,
    values_from = n
    ) %>%
  rename_with(~c("employed", "unemployed"), 2:3) %>%
  tidyr::replace_na(list(employed = 0, unemployed = 0)) %>%
  mutate(
    unemployment_rate = (unemployed / (unemployed + employed) * 100) %>% round(2),
    date              = paste0(api_ibge[[2]][[3]], " Q", api_ibge[[2]][[2]])
    ) %>%
  arrange(-unemployment_rate) %>%
  left_join(states, by = "states") %>%
  rename("woe-name" = states)


# Labor force occupation level
occupation_level <- raw_labor_force %>%
  select(
    date     = "Trimestre (C\u00f3digo)",
    variable = "Vari\u00e1vel",
    gender   = "Sexo",
    value    = "Valor"
    ) %>%
  filter(
    variable == "N\u00edvel de ocupa\u00e7\u00e3o, na semana de refer\u00eancia, das pessoas de 14 anos ou mais de idade"
    ) %>%
  slice_tail(n = 48) %>%
  mutate(
    date = stringr::str_replace(date, "(\\d{4})0(\\d{1}$)", "\\1 Q\\2"),
    variable = "Labor force occupation level",
    gender = recode(
      gender,
      "Homens" = "Man",
      "Mulheres" = "Woman"
      )
    )


# Labor force participation rate
participation_rate <- raw_labor_force %>%
  select(
    date     = "Trimestre (C\u00f3digo)",
    variable = "Vari\u00e1vel",
    gender   = "Sexo",
    value    = "Valor"
    ) %>%
  filter(
    variable == "Taxa de participa\u00e7\u00e3o na for\u00e7a de trabalho, na semana de refer\u00eancia, das pessoas de 14 anos ou mais de idade"
    ) %>%
  slice_tail(n = 48) %>%
  mutate(
    date = stringr::str_replace(date, "(\\d{4})0(\\d{1}$)", "\\1 Q\\2"),
    variable = "Labor force participation rate",
    gender = recode(
      gender,
      "Homens" = "Man",
      "Mulheres" = "Woman"
      )
    )


# Average real income from work, by age group
income <- raw_income %>%
  select(
    date     = "Trimestre (C\u00f3digo)",
    variable = "Vari\u00e1vel",
    age      = "Grupo de idade",
    value    = "Valor"
    ) %>%
  mutate(
    date = stringr::str_replace(date, "(\\d{4})0(\\d{1}$)", "\\1 Q\\2"),
    variable = "Average real income from work",
    id = rep(c("1","2","3","4","5","6"), length(date)/6),
    age = purrr::reduce2(
      .x    = c(" anos", " ou mais", " a "),
      .y    = c("", " or more", " to "),
      .init = age,
      stringr::str_replace
      )
    ) %>%
  tidyr::drop_na() %>%
  slice_tail(n = 78)


# Gini index by state
raw_pnadct_svy <- raw_pnadct %>%
  PNADcIBGE::pnadc_design() %>%
  convey::convey_prep()

gini <- survey::svyby(~VD4020, by = ~UF, raw_pnadct_svy, convey::svygini, na.rm = TRUE) %>%
  select("states" = UF, VD4020) %>% # Effective monthly income from all jobs
  left_join(states, by = "states") %>%
  as_tibble() %>%
  mutate(
    date  = paste0(raw_pnadct$Ano[1], " Q", raw_pnadct$Trimestre[1]),
    variable = "Gini index",
    value = round(VD4020, 2)
    ) %>%
  select(
    date,
    code,
    states,
    variable,
    value
    ) %>%
  arrange(states)




# Save data ---------------------------------------------------------------


if (0L %in% purrr::map_dbl(mget(ls()), length)) {

  stop("Some objects are zero in length.", call. = FALSE)

  } else
    {
    # Aggregate data
    rm(raw_pnadct_svy, envir = environment())
    imported_data_labor <- mget(ls(pattern = "raw_|api_"))


    # Remove unnecessary objects
    rm(
      list  = c(lsf.str(), ls(pattern = "raw_|api_")),
      envir = environment()
      )


    # Save RDATA file
    save(
      list  = ls(),
      file  = file.path(file.path("./inst/extdata"), "labor.Rdata"),
      envir = environment()
      )
    }

}