recode.R

## Labels have been manually reviewed
label <- read.csv("~/unhcr_r_project/displacement/data/label.csv")

## let's recode the variable of the dataset using short label
names(master) <- label[,3]
#names(master)
rm(label)

## Format date correctly
master$Date.Displacement <- as.Date(master$Date.Displacement, "%d-%m-%Y")
master$Current.Round.Date <- as.Date(master$Current.Round.Date, "%d %b %Y")

#master$Master.Families <- as.numeric(master$Master.Families)

master$Master.Families <- as.numeric(as.character(master$Master.Families))

master$PlaceID <- as.factor(master$PlaceID)
str(master)
#names(master)



# quick check on the total
master$total <- psum( master$IDPs.in.Camps.or.transit.camps,
                      master$Rented.Hotel,
                      master$Rented.House ,
                      master$IDP.Owned.House,
                      master$With.Relative ,
                      master$With.HC.non.Relative ,
                      master$School.Building,
                      master$Mosques.Holly.Shrines,
                      master$Abandoned.public.buildings.under.construction,
                      master$Collective.centres,
                      master$Informal.settlements,
                      master$Military.Camps,
                      master$Unknown.or.other,
                      na.rm=TRUE)*6

# adding month.displacement for further graph
master$Month.Displacement <- format(master$Date.Displacement,"%b-%y")
#levels(master$Month.Displacement)
master$Month.Displacement <- factor(master$Month.Displacement, levels = c("Dec-13","Jan-14","Feb-14","Mar-14", "Apr-14","May-14","Jun-14","Jul-14","Aug-14","Sep-14"))

## Adding some aggregation
master$communal.setting <- psum( master$IDPs.in.Camps.or.transit.camps,
                                 master$School.Building,
                                 master$Mosques.Holly.Shrines,
                                 master$Abandoned.public.buildings.under.construction,
                                 master$Collective.centres,
                                 master$Informal.settlements,
                                 master$Military.Camps,
                                 na.rm=TRUE)*6

master$private.setting <- psum( master$Rented.Hotel,
                                master$Rented.House ,
                                master$IDP.Owned.House,
                                master$With.Relative ,
                                master$With.HC.non.Relative,
                                na.rm=TRUE)*6

master$pc.communal <- master$communal.setting/master$total
master$pc.camp <- (master$IDPs.in.Camps.or.transit.camps*6)/master$total

# Recoding priority needs
#levels(master$priority.need.of.the.displaced.population.in.the.location)

## Parsing the variable and multiplying by total population
master$priority.Food <- as.numeric(with(master,
                                        ifelse(grepl("Food|Fis|FIand",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                     master$priority.need.of.the.displaced.population.in.the.location),
                                               paste0(1), 0)) ) * master$total

master$priority.Shelter <- as.numeric( with(master,
                                            ifelse(grepl("Shelter|Housing",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                         master$priority.need.of.the.displaced.population.in.the.location),
                                                   paste0(1), 0))) * master$total

master$priority.NFI <- as.numeric( with(master,
                                        ifelse(grepl("NFI|NFIs|non-food|CRIs",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                     master$priority.need.of.the.displaced.population.in.the.location),
                                               paste0(1), 0))) * master$total

master$priority.Water <- as.numeric( with(master,
                                          ifelse(grepl("water",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                       master$priority.need.of.the.displaced.population.in.the.location),
                                                 paste0(1), 0))) * master$total

master$priority.Cash <- as.numeric( with(master,
                                         ifelse(grepl("financial|Cash",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                      master$priority.need.of.the.displaced.population.in.the.location),
                                                paste0(1), 0))) * master$total

master$priority.Health <- as.numeric( with(master,
                                           ifelse(grepl("health|Healthy|Medical|Medicine|healthcare",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                        master$priority.need.of.the.displaced.population.in.the.location),
                                                  paste0(1), 0))) * master$total

master$priority.Child <- as.numeric( with(master,
                                          ifelse(grepl("children|Child|",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                       master$priority.need.of.the.displaced.population.in.the.location),
                                                 paste0(1), 0))) * master$total

#head(master)
#levels(master$Note)
#levels(master$assistance.previously.received)
#levels(master$sufficient.access.to.food)

#str(master)

#  names(master)

###################################
##### Size and type of displacement

# - # of IDP hosted in communal setting  -- a score can be given based on the size on the population
# a classification is computed on 5 classes using different algorythm

master$class.total <-findCols(classIntervals(master$total, n = 5, style = "jenks"))
#master$class.total <-findCols(classIntervals(master$total, n = 5, style = "quantile"))


# print(classIntervals(master$total, n = 5, style = "jenks"), unique=FALSE)
# pal1 = brewer.pal(5,"Greens")
# plot(classIntervals(master$total, n = 5, style = "jenks"), pal=pal1, main="Classification per total population")


# - % of IDP hosted in communal settings / private settings -- a score will be given based on this %

master$class.pccommunal <-findCols(classIntervals(master$pc.communal, n = 5, style = "jenks"))
# print(classIntervals(master$pc.communal, n = 5, style = "jenks"), unique=FALSE)


###################################
##### Impact on local population

# - a specific score could be given if idps are hosted in school building as this has a disruptive impact on local population
master$class.school <-findCols(classIntervals(master$School.Building, n = 5, style = "jenks"))


###################################
##### Exclusion criteria

# - a specific score could be given if idps are already in camp (compared to other communal setting)
master$class.camp <-findCols(classIntervals(master$IDPs.in.Camps.or.transit.camps, n = 5, style = "jenks"))

# - non access to services (in priority FOOD, WATER, NFI) -- a score can be given if no access is reported


# - specific score depending on the date of displacement (pre-june /post-june).
##master$class.date  <-findCols(classIntervals(as.numeric(master$Date.Displacement), n = 5, style = "jenks"))

master$class.date  <-as.numeric(master$Date.Displacement)

master$class.date  <-  as.numeric( with(master,
                                        ifelse(as.numeric(master$Date.Displacement)> as.numeric(as.Date('01-06-2014', "%d-%m-%Y")) ,
                                               paste0(0), 1)))
## New label to define wether the 
master$datecut  <-  with(master,
                         ifelse(as.numeric(master$Date.Displacement)> as.numeric(as.Date('01-06-2014', "%d-%m-%Y")) ,
                                paste0("Post-June"), "Pre-June"))
## New label to define wether the 
master$datecut <- factor(master$datecut, levels = c("Pre-June","Post-June"))


# - specific flag if the location has previously received assistance
#levels(master$has.received.assistance)
master$Assistance.received <- as.numeric( with(master,
                                               ifelse(grepl("Yes",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                            master$has.received.assistance),
                                                      paste0(0), 1)))

# - specific flag when shelter has been defined as one of the priority need.

master$Shelterpriority <- as.numeric( with(master,
                                           ifelse(grepl("Shelter|Housing",ignore.case = TRUE,fixed = FALSE, useBytes = FALSE,
                                                        master$priority.need.of.the.displaced.population.in.the.location),
                                                  paste0(1), 0)))


## identify locations where refugee camp already exist in the next 20kms!!!!


## Try to find place with same name but different master id
#extractplacecountplace <- aggregate(cbind(total) ~ Place ,  data = master, FUN = sum, na.rm = TRUE)
#extractplacecountmasterid <- aggregate(cbind(total) ~ Master.ID ,  data = master, FUN = sum, na.rm = TRUE)


####

## Multiplying the family number by the average
master$Unknown.or.other <- master$Unknown.or.other*6
master$IDPs.in.Camps.or.transit.camps <- master$IDPs.in.Camps.or.transit.camps*6
master$Military.Camps <- master$Military.Camps*6
master$Informal.settlements <- master$Informal.settlements*6
master$Collective.centres <- master$Collective.centres*6
master$Abandoned.public.buildings.under.construction <- master$Abandoned.public.buildings.under.construction*6
master$Mosques.Holly.Shrines <- master$Mosques.Holly.Shrines*6
master$School.Building <- master$School.Building*6
master$With.HC.non.Relative <- master$With.HC.non.Relative*6
master$With.Relative <- master$With.Relative*6
master$IDP.Owned.House <- master$IDP.Owned.House*6
master$Rented.House <- master$Rented.House*6
master$Rented.Hotel <- master$Rented.Hotel*6

#Using the following reclassification

#IDPs.in.Camps.or.transit.camps ;  IDPs.in.Camps.or.transit.camps
#Rented.Hotel ;  Rented.or.Owned.Accomodation
#Rented.House ;   Rented.or.Owned.Accomodation
#IDP.Owned.House ;   Rented.or.Owned.Accomodation
#With.Relative ;   Hosted.Accomodation
#With.HC.non.Relative ; Hosted.Accomodation
#School.Building ; Self-settled.in.public.building
#Mosques.Holly.Shrines  ; Self-settled.in.public.building
#Abandoned.public.buildings.under.construction ; Self-settled.in.public.building
#Collective.centres ; Collective.centres
#Informal.settlements  ; Informal.settlements
#Military.Camps ; Informal.settlements
#Unknown.or.other  ; Unknown.or.other

master$Hosted.or.Owned.Accomodation <- psum( master$IDP.Owned.House,
                                             master$With.Relative,
                                             master$With.HC.non.Relative, 
                                             na.rm=TRUE)

master$Rented.Accomodation <- psum(master$Rented.House,
                                   master$Rented.Hotel,
                                   na.rm=TRUE)

master$Organised.site   <- psum( master$Collective.centres,
                                 master$Mosques.Holly.Shrines,
                                 master$Military.Camps,  
                                 na.rm=TRUE)

master$Improvised.site   <- psum( master$IDPs.in.Camps.or.transit.camps,
                                  master$Abandoned.public.buildings.under.construction,
                                  na.rm=TRUE)


master$Squatted.schools <- psum( master$School.Building,
                                 na.rm=TRUE)

master$Open.air   <- psum( master$Informal.settlements,
                           na.rm=TRUE)

master$base2  <- "Rest of Iraq"
master$base2[master$Governorate=="Dahuk"]  <- "KRI"
master$base2[master$Governorate=="Erbil"]  <- "KRI"
master$base2[master$Governorate=="Sulaymaniyah"]  <- "KRI"

master$base2[master$District=="Khanaqin"]  <- "Disputed Territories"
master$base2[master$District=="Akre"]  <- "Disputed Territories"
master$base2[master$District=="Al-Shikhan"]  <- "Disputed Territories"
master$base2[master$District=="Kifri"]  <- "Disputed Territories"
#master$base2[grepl("Balance MoP", master$Place) ]  <-"Balance MoP"
#master$base2[is.na(master$base) ]  <-"-"
master$base2 <- as.factor(master$base2)




## Adding some aggregation

master$IDPs.in.Camps.or.transit.camps10 <- as.numeric( with(master, ifelse((master$IDPs.in.Camps.or.transit.camps>60) ,
                                                                           paste0(master$IDPs.in.Camps.or.transit.camps), 0)))
master$School.Building10 <- as.numeric( with(master, ifelse((master$School.Building>60) ,
                                                            paste0(master$School.Building), 0)))
master$Mosques.Holly.Shrines10 <- as.numeric( with(master, ifelse((master$Mosques.Holly.Shrines>60) ,
                                                                  paste0(master$Mosques.Holly.Shrines), 0)))
master$Abandoned10 <- as.numeric( with(master, ifelse((master$Abandoned.public.buildings.under.construction>60) ,
                                                      paste0(master$Abandoned.public.buildings.under.construction), 0)))
master$Collective.centres10 <- as.numeric( with(master, ifelse((master$Collective.centres>60) ,
                                                               paste0(master$Collective.centres), 0)))
master$Informal.settlements10 <- as.numeric( with(master, ifelse((master$Informal.settlements>60) ,
                                                                 paste0(master$Informal.settlements), 0)))
master$Military.Camps10 <- as.numeric( with(master, ifelse((master$Military.Camps>60) ,
                                                           paste0(master$Military.Camps), 0)))


master$communal.setting10 <- psum( master$IDPs.in.Camps.or.transit.camps10,
                                   master$School.Building10,
                                   master$Mosques.Holly.Shrines10,
                                   master$Abandoned10,
                                   master$Collective.centres10,
                                   master$Informal.settlements10,
                                   master$Military.Camps10,
                                   na.rm=TRUE)




write.csv(master, '~/unhcr_r_project/displacement/out/data/master.csv', row.names=TRUE)