Functions for figuring out where user specified thresholds are met in a vector using R

ScopingFunctions.R

#######################################
# Example data
######################################

testdata<-replicate(10, rpois(100, 20)) 



###########################################
# This function tells us how far we have to
# go before reaching a cutoff in a variable
# by sorting the vector, then finding how far 
# to go. Note that the cutoff is expressed in 
# percentage terms
############################################

fixcumsum<-function(x,cutoff){
  x<-x[order(-x)] #sort vector descending
  xb<-cumsum(x) # take cumulative sum
  xc<-xb/sum(x,na.rm=T) #express proportionally
  length(xc[xc<cutoff]) #count number of items until 
  #                     threshhold is exceeded
}

##########################################
# This function allows us to see what % 
# of observations are found at a given threshold
# again expressed in percentage terms.
############################################

cutoff<-function(x,thresh){
  #x is the column or variable
  #thresh is the number to count to 
  x<-x[order(-x)] # sort vector descending
  xb<-cumsum(x)   # take cumulative sum (order matter)
  xc<-xb/sum(x,na.rm=T) # express proportionally
  xc[thresh]            # report cumulative percentage at given threshold
}

###############################
# Now we have to simply apply
# over a data element such as a 
# matrix or a dataframe
###############################

#here we apply the object to columns we specify that apply 
#(data, 2, function, function var) 
#where 2 tells R to go column-wise, which is appropriate for 
# this data shape, and we specify function parameters after we 
# tell R the function

thresh3<-apply(testdata,2,cutoff,thresh=3)
thresh5<-apply(testdata,2,cutoff,thresh=5)
thresh10<-apply(testdata,2,cutoff,thresh=10)

# we store these as variables because a vector is produced that 
# shows us the value for each function that was applied to 


cutoffs<-cbind(thresh3,thresh5,thresh10)

# We combine these into a data frame

# We make them into % for Excel purposes (not necessary)
cutoffs<-cutoffs*100