assumptions.R

#
# Assumptions and parameters used for constructing the TCI index
#


# minimum number of indicators a country must have to be included

min.num.inds <- 6

# chosen indicators

indicators <- list(

  # Transboundary Water
  "Transboundary_Water" = list( fname = "transboundary_water.csv"),

  # Bilateral climate weighted FDI
  "FDI" = list (fname = "fdi.csv"),
  
  # Remittances
  "Remittances" = list (fname = "remittances.csv"),

  # Openness to asylum  
  "Asylum" = list (fname = "refugees.csv"),
  
  # Migration from climate vulnerable countries
  "Migration" = list (fname = "migration.csv"),
  
  # Trade openness
  "Trade_Openness" = list (fname = "trade_openness.csv"),
  
  # Cereal import dependency
  "Cereal_Imports" = list (fname = "cereal.csv", title = "Cereal Import Dependency"),

  # Embedded water risk
  "Embedded_Water" = list (fname = "embedded_water.csv"),

  # KOF Globalization Index
  "Globalization" = list (fname = "kof_globalization.csv")

)

# number of quantile 'bins' to use for generating indicator scores
num.quantiles <- 10

# Ranking function to calculate indicator scores
 score <- function( values , bins ){
    
    # apply a raw ranking of values
    rnk <- rank(values , ties = 'min', na.last = 'keep')
    
    # note: The ties = "min" parameter ensures that countries with the same
    #       value receive the same rank.
    #       The na.last='keep' maintains missing values in their proper place.
    
    # bin boundaries
    bin.bounds = quantile(unique(rnk), probs = 0:bins/bins, na.rm= TRUE)
    
    #return a vector of scores
    cut( rnk, bin.bounds , include.lowest =TRUE, labels = FALSE)
  
  }

#####

# Sensitivity analysis

# possible noise to add, as a percentage of range
percent.error = c(15)



###########################
# Alternative scoring function


score2 <- function(values, bins){
    
    set.seed(1989)
    
    # apply a raw ranking of values
    rnk <- rank(values , ties = 'min', na.last = 'keep')

    # omit (and record position of) NA values
    rnk <- na.omit(rnk)

    # note: The ties = "min" parameter ensures that countries with the same
    #       value receive the same rank.
    #       The na.last='keep' maintains missing values in their proper place.
    
    # apply kmeans classification to rank levels to deal with duplicate values 
    # K-means will find groupings which minimize 
    clu <- kmeans(rnk, bins)$cluster
    
    
    
    #reassign cluster ids starting with the smallest group

    a <- aggregate(rnk, by = list(clu), FUN=min, na.rm=TRUE)
    out <- rank(a$x)[clu]
    
    for (i in attr(rnk, 'na.action')){
    
     out <- append(out, NA, after = i-1 )
    
    }
    
    out
}