require('tidyverse')
require('magrittr')

# import the data
tall_index <- as_tibble(read.csv('tall_index.csv', fileEncoding="UTF-8-BOM"))
grouping <- as_tibble(read.csv('grouping04.csv', fileEncoding="UTF-8-BOM"))

# normalize the levels of the vectors for join
combined <- sort(union(levels(tall_index$Country), levels(grouping$Country))) 
#create the joined file, filter our the countries that do not have a Level attached
combined_tall_index <-tall_index %>%
  left_join(mutate(grouping, Country=factor(Country, levels=combined))) %>%
  filter(!is.na(Level))

#calculate the two averages
countries_averages <- combined_tall_index %>% group_by(Country)%>%  
  summarize(country_average = mean(fragile_states_index), Level= first(Level))
levels_averages <- combined_tall_index %>% group_by(Level)%>%  
  summarize(level_average = mean(fragile_states_index))
# join back the table
new_tall <- left_join(countries_averages, levels_averages)

#add the new variable containing the difference
new_tall %<>% mutate(difference= country_average - level_average)

#write the countries list with a superior average over its level
new_tall %>% 
  filter(difference >= 0) %>% 
  write_csv("positives_R.csv")

# wirte the countries list with an average below its level group
new_tall %>% 
  filter(difference < 0) %>% 
  write_csv("negatives_R.csv")