## set your own working directory
setwd("/Users/rhwiley/Desktop")

## use revised data file
MbbsAll <- read.table(file="MBBS_ALLCOUNTIES_REV.txt", header=TRUE, sep="", 
	quote="", dec=".", 
	na.strings="X", colClasses=NA, nrows=-1, skip=0, 
	check.names=TRUE, strip.white=TRUE, blank.lines.skip=TRUE)

names(MbbsAll)


############################################


# calculate values for each species for missing route-years
# first create a new data.frame that excludes all years with missing routes (zzz)
# use this data.frame to calculate the mean numbers for each species in years with data
# substitute these means for each species in the original data.frame (MbbsAll) 

#### SET THE COUNTY FOR THIS PROCEDURE ####
County = "CHATHAM"
RoutesWithMissingData = NULL
NumRoutesWithMissingData = 0

# create a new data.frame for this county
### DONT FORGET THE COMMAS WHEN SELECTING COLUMNS OR ROWS!!
zzz <- MbbsAll[MbbsAll$COUNTY==County, ]

# then prune this data.frame to eliminate years with missing data
# number of rows in MbbsAll = 394
for (ii in 2:394)
{
	# if this row is in the correct county, get $DATE2 (column 6) from this row	
	if(MbbsAll[MbbsAll$ROWNUM==ii,]$COUNTY == County) 
	{
		xxx = MbbsAll[ii, 6]  ## xxx = MbbsAll[MbbsAll$ROWNUM==ii,]$DATE2
	}
	# if the date (hence all data) is missing, add the route, year, and row number in MbbsAll to a list
	if(is.na(xxx)) 
	{
		## yyy = TRUE
		YearWithMissingData = MbbsAll[MbbsAll$ROWNUM==ii, ]$YEAR
		RoutesWithMissingData = c(RoutesWithMissingData, MbbsAll[MbbsAll$ROWNUM==ii, ]$ROUTE)
		RoutesWithMissingData = c(RoutesWithMissingData, YearWithMissingData, ii)
		NumRoutesWithMissingData = NumRoutesWithMissingData + 1
		
		# now rewrite the pruned data.frame without this year
		zzz = zzz[zzz$COUNTY==County & zzz$YEAR!=YearWithMissingData, ]
	}
}

# some checks on progress so far
summary(zzz[50])
summary(zzz$ACFL)
mean(zzz[50])
round(mean(zzz[50]))
unique(zzz$YEAR)
RoutesWithMissingData
NumRoutesWithMissingData

# following line gives an error when used for an assignment in the loops below !!!
mean(zzz[zzz$ROUTE==as.character(RoutesWithMissingData[1]), ][120])
mean(zzz[120]) # mean over all routes

# use the list of missing routes and the pruned data.frame to substitute values in MbbsAll
for (jj in 1:NumRoutesWithMissingData)
{
	# rr is an index to the row numbers with missing data in MbbsAll (in the list RoutesWithMissingData) 
	rr=(jj-1)*3
	# get a route with missing data
	MissingRoute = RoutesWithMissingData[rr+1]  # MissingRoute = as.character(RoutesWithMissingData[rr+1])

	for (kk in 13:120)
	{
		# means for this route in years without missing data (species are in columns 13-120)
		RouteMeanColumnKk = mean(zzz[zzz$ROUTE==MissingRoute, ][kk])
		RouteMeanDate = mean(zzz[zzz$ROUTE==MissingRoute, ]$DATE2)
		# substitute these means in the appropriate lines of MbbsAll
		MbbsAll[RoutesWithMissingData[rr+3], kk] <- round(RouteMeanColumnKk)
		MbbsAll[MbbsAll$ROWNUM==RoutesWithMissingData[rr+3], ]$DATE2 <- round(RouteMeanDate)
	}
}



############################################




## extract the data for one county at a time
# County = "ORANGE"
MbbsCountyTotals <- MbbsAll[MbbsAll$COUNTY==County & MbbsAll$ROUTE=="TOTALSP", ]

## example of summary stats for Carolina Wren (CARW)
x <- MbbsCountyTotals $CARW

summary(x)

boxplot.stats(x, coef = 1.5, do.conf = TRUE, do.out = TRUE)


############################################


## plot graphs for each species ## with line color and line type  ## better comments
## prints ten pdfs in one file
## quartz(width=9, height=12) ## for use on-screen -- delete pdf below -- also enable break below

j = 1
TotalSpecies = 120
ListToPlot=NULL ##do not use list(NULL)

for (i in 13:TotalSpecies+1)
## column 13 has the annual totals for the first species of bird, column 118 has the totals for the last species 
{
	if (j <= 10 && i <= TotalSpecies)
	# get a list of ten species to plot
	{
		# only include species with a mean annual number of individuals > 5
		if (summary(MbbsCountyTotals[[i]])[[4]]>5)
		# if mean>5 add the column number to the list to plot
		{
			ListToPlot = c(ListToPlot, i)
			j=j+1
		}
	}

	else
	# plot these ten species and reset the list
	{
		NumberToPlot = j-1
		# following data.frame is not used
		MbbsCountySub <- subset(MbbsCountyTotals, 
			ROUTE == "TOTALSP" & COUNTY == County,
			select = ListToPlot)

		if(i > TotalSpecies) LastNumber = TotalSpecies else LastNumber = i
		Filename = paste("/Users/rhwiley/Desktop/", County, LastNumber, 
			".pdf", sep="")
		pdf(file=Filename, width=9, height=12, onefile=TRUE)

		# two columns of five graphs
		layout(
			matrix(c(1,2,3,4,5,6,7,8,9,10),
				ncol=2, nrow=5, byrow=FALSE),
			widths=rep(lcm(10),2),
			heights=c(rep(lcm(5),5)),
			respect=FALSE	
		)
		par(
			mar=c(4,4,0,2), # (2,4,2,4) B,L,T,R
			oma=c(2,2,4,1), # (2,2,1,1)
			mgp=c(3,1,0) # margin line for y axis labels, x axis labels, axis lines (3,1,0)
		)

		for (k in 1:NumberToPlot)
		{	
			Ltpk=ListToPlot[k]
			YLabel = names(MbbsCountyTotals)[Ltpk] 
			if(YLabel=="TOTALRT") YLabel="ALL SPECIES"
			plot(
				MbbsCountyTotals[[Ltpk]] ~ MbbsCountyTotals$YEAR, type="o",
				xlab = "", ylab=YLabel,
				ylim=c(0,10*((max(MbbsCountyTotals[[Ltpk]])%/%10)+1)), 
				#log="y",
				#main="MINI BREEDING BIRD SURVEY",
				#sub=paste(County, " COUNTY, NC"),
				main=""
			)

			# calculate the regression line
			Ltpk_lm = lm(MbbsCountyTotals[[Ltpk]] ~ MbbsCountyTotals$YEAR)
			Ltpk_lm_summ <- summary(Ltpk_lm)

			# get the P level for the regression			
			if(Ltpk_lm_summ$coefficients[2,4]>0.1) Psymbol="X";
			if(Ltpk_lm_summ$coefficients[2,4]<=0.1 & Ltpk_lm_summ$coefficients[2,4]>0.05) Psymbol="";
			if(Ltpk_lm_summ$coefficients[2,4]<=0.05 & Ltpk_lm_summ$coefficients[2,4]>0.01) Psymbol="*";
			if(Ltpk_lm_summ $coefficients[2,4]<=0.01) Psymbol="**";

			# if P<0.1 plot the line, if P<0.05 show the significance level
			# adjust line type and width to reflect P level
			# lty = 2 for dashed line, lwd = 3 for 3x wider than default
			if (Psymbol!="X")
			{	

				# default values
				LineType=1
				LineWidth=1
				LineColor="black"

				# lty=2 for dashed line
				if(Psymbol == "") LineType = 2

				# colored lines are too much!	
#				else 
#				{
#					if(Ltpk_lm_summ$coefficients[2,1]<0) LineColor ="red"
#					else LineColor ="green"
#				}

				# lwd=3 for 3x wider than default	
				if(Psymbol =="**") LineWidth =3
				
				abline(Ltpk_lm, lty=LineType, lwd=LineWidth, col=LineColor)
				ycoord=10*((max(MbbsCountyTotals[[Ltpk]])%/%10)+0.5)
				text(2008, ycoord, Psymbol, font=2, cex=3) 
				##font=2 is bold, cex=2 multiplies size by 2
				##coordinates for text are the coordinates of the plot itself!!!
			}
		}
		title(line=1, main=paste(County, " County\nMini Breeding Bird Survey"), cex.main=1.5,
			outer=TRUE)

		# break # if you need to do some de-bugging!
		dev.off()

		# clear list of species to plot and reset j
		ListToPlot=NULL
		j = 1
	}
}