setwd("/home/suhel/Documents/MigrantWatch/AnalysisScripts/PiedCuckoo/Analysis2011")

## first explore all PC data and filter those for 2009 and 2010
dat <- read.csv("all_pc_records_30April2011.csv")
## correct some city names
dat$City[dat$City=="Dehra Dun"] <- "Dehradun"
dat$City[dat$City=="Kanpur City"] <- "Kanpur"

## add a serial number
dat$Sno <- 1:nrow(dat)
dat$Date <- as.Date(dat$Date)
## remove places south of 15 degrees North
dat <- subset(dat, Latitude>=15)
## Could also remove first visits, but not doing so for now
# dat <- subset(dat, Observation.frequency!="First visit")
## choose appropriate date range
dat.2009 <- subset(dat, Date>=as.Date("2009-05-01") & Date <=as.Date("2009-07-15"))
dat.2010 <- subset(dat, Date>=as.Date("2010-05-01") & Date <=as.Date("2010-07-15"))
## remove extra City names
dat.2009$City <- factor(dat.2009$City)
dat.2010$City <- factor(dat.2010$City)
## order sightings in ascending order within City
dat.2009 <- dat.2009[order(dat.2009$City, dat.2009$Date),]
dat.2010 <- dat.2010[order(dat.2010$City, dat.2010$Date),]
## take earliest sighting for each City
dat.2009 <- dat.2009[with(dat.2009, match(unique(City), City)) ,]
dat.2010 <- dat.2010[with(dat.2010, match(unique(City), City)), ]

## read monsoon arrival data
mon <- read.csv("monsoon_arrival_0910.csv")
mon$Monsoon_arrival <- as.Date(mon$Monsoon_arrival)
mon.2009 <- subset(mon, Year==2009)
mon.2010 <- subset(mon, Year==2010)
## take only unique cities
mon.2009 <- mon.2009[match(unique(mon.2009$City), mon.2009$City),]
mon.2010 <- mon.2010[match(unique(mon.2010$City), mon.2010$City),]
## match cities to fill in monsoon dates
dat.2009$mon.arr <- mon.2009[match(dat.2009$City, mon.2009$City), "Monsoon_arrival"]
dat.2010$mon.arr <- mon.2010[match(dat.2010$City, mon.2010$City), "Monsoon_arrival"]
## check arrival dates against monsoon master file.
dat.2009[,c("City", "Date", "mon.arr")]
dat.2010[,c("City", "Date", "mon.arr")]
## missing monsoon data
subset(dat.2009, is.na(mon.arr), select=c("City", "State"))
subset(dat.2010, is.na(mon.arr), select=c("City", "State"))
  ## no missing data

## which locations have data from both years
bothyrs <- intersect(dat.2009$City, dat.2010$City)
bothyrs <- data.frame(City=bothyrs, No=1:length(bothyrs))
bothyrs.2009 <- data.frame(bothyrs, dat.2009[match(bothyrs$City, dat.2009$City), c("Date", "mon.arr")])
bothyrs.2010 <- data.frame(bothyrs, dat.2010[match(bothyrs$City, dat.2010$City), c("Date", "mon.arr")])
names(bothyrs.2009)[3:4] <- c("cuck09", "mon09")
names(bothyrs.2010)[3:4] <- c("cuck10", "mon10")

## Now plot pied cuckoo date versus monsoon arrival date separately for 2009 and 2010
svg("PCblog-May2011-plot1.svg", width=12, height=5) ## if file is to be printed
par(mfrow=c(1,2), mar=c(4,5,2,2), las=1)
plot(Date ~ mon.arr, data=dat.2009, pch=16, xlim=as.Date(c("2009-05-15", "2009-07-05")), ylim=as.Date(c("2009-05-15", "2009-07-15")), xaxt="n", yaxt="n", xlab="Monsoon arrival", ylab="First sighting")
text(bothyrs.2009$mon09-1, bothyrs.2009$cuck09+1, bothyrs.2009$No)
axis(side=1, at=as.Date(c("2009-05-15", "2009-06-01", "2009-06-15", "2009-07-01")), labels=c("15 May", "1 June", "15 June", "1 July"))
axis(side=2, at=as.Date(c("2009-05-15", "2009-06-01", "2009-06-15", "2009-07-01", "2009-07-15")), labels=c("15 May", "1 June", "15 June", "1 July", "15 July"))
abline(c(0,1)); abline(c(-5, 1), lty=2); abline(c(-30, 1), lty=3)
plot(Date ~ mon.arr, data=dat.2010, pch=16, xlim=as.Date(c("2010-05-15", "2010-07-05")), ylim=as.Date(c("2010-05-15", "2010-07-15")), xaxt="n", yaxt="n", xlab="Monsoon arrival", ylab="First sighting")
text(bothyrs.2010$mon10-1, bothyrs.2010$cuck10+1, bothyrs.2010$No)
axis(side=1, at=as.Date(c("2010-05-15", "2010-06-01", "2010-06-15", "2010-07-01")), labels=c("15 May", "1 June", "15 June", "1 July"))
axis(side=2, at=as.Date(c("2010-05-15", "2010-06-01", "2010-06-15", "2010-07-01", "2010-07-15")), labels=c("15 May", "1 June", "15 June", "1 July", "15 July"))
abline(c(0,1)); abline(c(-5, 1), lty=2); abline(c(-30, 1), lty=3)
legend("topleft", legend=c("Same day as monsoon arrival",
    "5 days before monsoon arrival",
    "30 days before monsoon arrival"),
    lty=1:3, bty="n") 
dev.off() ## if file is to be printed

## plot the 12 common locations separately
both <- data.frame(bothyrs.2009, bothyrs.2010[,c("cuck10", "mon10")])
  ## assign a common year, for the purpose of plotting
both$cuck09.com <- as.Date(paste("2011", substr(both$cuck09, 6, 10), sep="-"))
both$cuck10.com <- as.Date(paste("2011", substr(both$cuck10, 6, 10), sep="-"))
both$mon09.com <- as.Date(paste("2011", substr(both$mon09, 6, 10), sep="-"))
both$mon10.com <- as.Date(paste("2011", substr(both$mon10, 6, 10), sep="-"))

## Plot monsoon on x, sighting on y
svg("PCblog-May2011-plot2.svg", width=5, height=5) ## if plot to be printed to file
par(mfrow=c(1,1), las=1)
with(both, plot(mon09.com, cuck09.com, 
  xlim=as.Date(c("2011-05-15", "2011-07-15")),
  ylim=as.Date(c("2011-05-15", "2011-07-15")), pch=16,
  xlab="Monsoon arrival",ylab="First sighting", xaxt="n", yaxt="n"))
axis(side=1, at=as.Date(c("2011-05-15", "2011-06-01", "2011-06-15", "2011-07-01", "2011-07-15")), labels=c("15 May", "1 June", "15 June", "1 July", "15 July"))
axis(side=2, at=as.Date(c("2011-05-15", "2011-06-01", "2011-06-15", "2011-07-01", "2011-07-15")), labels=c("15 May", "1 June", "15 June", "1 July", "15 July"))
with(both, points(mon10.com, cuck10.com, pch=16, col="red"))
with(both, text(mon09.com-1, cuck09.com+1, No))
abline(c(0, 1), lty=2)
for(i in 1:nrow(both)){
  with(both[i,], lines(c(mon09.com, mon10.com), c(cuck09.com, cuck10.com)))
}
legend("bottomright", legend=c(2009, 2010), pch=16, col=c("black", "red"), bty="n")
## Add hypothetical location where changed in cuckoo arrival matches change in monsoon
points(as.Date(c("2011-06-01", "2011-06-11")), as.Date(c("2011-05-20", "2011-05-30")), pch="*", cex=3, col=1:2)
lines(as.Date(c("2011-06-01", "2011-06-11")), as.Date(c("2011-05-20", "2011-05-30")), lty=2, lwd=1.5)
text(as.Date("2011-05-31"), as.Date("2011-05-18"), labels="Example", cex=0.8)
dev.off() ## end print

## Output processed text files for others to use
dat.2009$Year <- 2009
dat.2010$Year <- 2010
dat.complete <- rbind(dat.2009, dat.2010)
dat.complete <- subset(dat.complete, select=c(Species, Location.name,
  City, State, Reporter, Date, Sighting.type, On.behalf.of, Latitude,
  Longitude, mon.arr) )
names(dat.complete)[names(dat.complete)=="Date"] <- "Sighting.date"
names(dat.complete)[names(dat.complete)=="mon.arr"] <- "Monsoon.arrival"

both.years <- both[,1:6]
names(both.years)[-(1:2)] <- c("Sighting.date.09", "Monsoon.09",
    "Sighting.date.10", "Monsoon.10")
    
## Output these data.frames
write.csv(dat.complete, file="PCblog-May2011-Plot1Data.csv",row.names=FALSE)
write.csv(both.years, file="PCblog-Mat2011-Plot2Data.csv", row.names=FALSE)