DewarManual.R

#################################################################
#
# R-Studio @ Macquarie University, Sydney, Australia
# 15.05.2017 - Adela Sobotkova
#
# Topic: Dealing with Contemporaneity in Survey data through a Census approach (a la Dewar 1991)
#
################################################################

#### Set Workspace, if your data is tabular

#setwd("D:/Users/MQ20149304/Documents/RStudio/Workshop1")
# to be continued >> Work in Progress, ask Adela about DewarUpgrade 

###If no workspace needed, enter parameters directly here (currently Shepelah region in Nitsan's dataset)

phase=207  # length in years of $ phase
vill=32 #product of a+b in Dewar
settle=39# (c+d)/phaselength, site establishment rate * 1000 (to match random integer data generated by sample function)
aban=58 # (a+d)/phaselength,site abandonment rate * 1000 (to match random integer data generated by sample function)
year=1 #starting year of the phase iteration

### Create a matrix in which you will aggregate the interim data
phase.results<-matrix(year, nrow = 0, ncol = 3,byrow=T)  
dimnames(phase.results)<-list(
  c(),  # row names
  c("Year", "CurrentVillages", "SD")) # column names 
print(phase.results)  # just a test message that this part of code is not broken, it shoud only print the column names.

### Create a loop that models the emergence of new villages on the basis of the existing site abandonment and creation rates

## The loop and random number generation
while(year<=phase){iter <- sample(1:1000,999,rep=T) # creates 1000 random numbers for every year of the phase, runs 999 times
iter2 <- ifelse(iter<=settle,1,0) # checks if the random number is below the settlement rate and if so, output 1, signalling that a settlement was newly inhabited
iter3 <- ifelse(iter<=aban,1,0) # checks if the random number is below the abandonment rate and if so, output 1, signalling that a village was abandoned
new<-iter2+vill-iter3  # adds the newly settled and substracts the abandoned settlements
vill<-mean(new)  # calculates the mean of the new villages in a given year on the basis of 999 iterations
x<-sd(new)   # calculates the standard deviation of new villages in a given year
year.results<-c(year,vill,x) #creates a vector of data from the annual statistic
phase.results<-rbind(phase.results,c(year,vill,x)) #combines all the individual annual iter into a table for the phase
year<-year+1
}

### Results #####

Occ<-mean(phase.results[,2])  # number of simultaneously occupied villages
Occ.sd<-sd(phase.results[,2])  # the standard deviation in the simultaneously occupied villages
Use.span<-1/(aban*0.001/Occ)  # mean use span of the villages in this phase. the multiplication reconverts the abandonment rate to original real number
Low.span<-1/(aban*0.001/(Occ-Occ.sd))   # mean use span minus 1 sd. the multiplication reconverts the abandonment rate to original real number
Hi.span<-1/(aban*0.001/(Occ+Occ.sd))  # mean use span plus 1 sd. the multiplication reconverts the abandonment rate to original real number
Span.err<- (Hi.span-Low.span)/2  # the standard deviation in the mean use span


### Final Messages #####

message("During this phase the number of simultaneously occupied villages is ca ",Occ,"+/-",Occ.sd)
message("The mean use span with 1 standard deviation is ",Use.span,"+/-",Span.err)

message("Less one sd the span estimate is ",Low.span)
message("Plus one sd the span estimate is ",Hi.span)


##### End of the Exercise ####################