# Inequality Key Figures - replication with R and `lissyrtools`

library(lissyrtools)
library(tidyverse)


# 0) Load Data and Prepare it 
data <- lissyuse(
  data = "lu",   # Examples: `data = c("de", "it")`; `data = c("it20", "it16")`; Run: get_countries_lis() to check other iso2 codes, or get_years_lis(c("it", "es")) to check which years are available.
  vars = c("dhi", "nhhmem65", "nhhmem17","sex", "relation", "hpartner")
) %>%
  map(~ .x %>% 
        filter(!is.na(dhi)) %>%
        filter(relation == 1000) %>%
        mutate(
          new_wgt = hwgt*nhhmem, 
          new_chld_wgt = hwgt * nhhmem17,
          new_eldr_wgt = hwgt * nhhmem65 
        )
  ) %>%
  apply_iqr_top_bottom_coding("dhi", "hwgt", type = "type_4") %>% 
  apply_sqrt_equivalisation("dhi")


medians <- map(
  run_weighted_percentiles(data, "dhi", "new_wgt", type = "type_4"), 
  "50%"
)





# 4) Poverty Rates for Children by Family Type (p-level originally as we need `sex` and `relation`)

poverty_line <- map(medians  , ~ .x  * 0.5)

data_with_pov_line <- map2(
  data,
  poverty_line,
  ~ .x %>%
    mutate(
      line = .y,
      poor = if_else(dhi < line, 100, 0)
    )
)


# Two-parent families (50%)

chld_pov_rate_2_parent <- data_with_pov_line %>% 
  map( ~ .x %>%
         filter(hpartner==1 & nhhmem17>0 & nhhmem>2)
  ) %>% 
  run_weighted_mean("poor", "new_chld_wgt")

# Single-mother families (50%)

chld_pov_rate_sing_mum <- data_with_pov_line %>% 
  map( ~ .x %>%
         filter(hpartner==0 & sex==2 & nhhmem17>0)
  ) %>% 
  run_weighted_mean("poor", "new_chld_wgt")

# Percentage living in Single-mother households 

chld_percent_sing_mum <- data_with_pov_line %>% 
  map( ~ .x %>%
         mutate(kidsm = if_else(sex==2 & hpartner==0 & nhhmem17>0, 100, 0, missing = 0))
  ) %>% 
  run_weighted_mean("kidsm", "new_chld_wgt")



# Collect them into a list
all_stats <- list(
  poortp = chld_pov_rate_2_parent,
  poorsm = chld_pov_rate_sing_mum,
  pkidsm = chld_percent_sing_mum
)

results <- map(all_stats,  ~ structure_to_plot(.x, print_columns = FALSE)) 

final <- imap(results , ~ .x %>% mutate(short_ikf = .y)) %>% 
  bind_rows() %>% 
  arrange(short_ikf, dname) %>%  
  mutate(
    indicator = case_when(
      short_ikf == "poortp"     ~ "Children Poverty Rates - Two-Parent Families (50%)",
      short_ikf == "poorsm"     ~ "Children Poverty Rates - Single-Mother Families (50%)",
      short_ikf == "pkidsm"     ~ "% Children Living in Single-Mother Families"
      )
  ) %>%
  rename(value_ikf_R = value)

write.csv(final)