## Getting the data in order; work with raw data files with tree and plot data, and combine into summarized data by plot (point).  

```{r}
#plot-level files. All the hot spots, and the subset of plots that were measured. Put together.

wd <- ('C:/Users/aabarker/Dropbox (Personal)/GypsyMothOak-Main/Hotspot2017LandsatValidation/HotspotDataArchive')
setwd(wd)
library(plyr)
latlong<-read.csv('LatLong.csv')
measplot<-read.csv('MeasuredPoints.csv')
spotplots<-merge(measplot,latlong, by = c('hotspot', 'point'), all.x=TRUE)

#tree-level measurements file.
trees<-read.csv('Trees.csv', header = TRUE)
str(trees)

treep<-merge(trees, spotplots, by=c('hotspot', 'point'), all.x=TRUE, all.y=TRUE)
#trees with exceedingly large dbh
#TooBig<-subset(treep, treep$DBH>40) #These 4 trees actually all check out ok with the data sheets. They are just big.

#remove NA lines except for the 3 plots that have no trees
#how many are there?
treepNA<-subset(treep, is.na(treep$DBH)) 
#remove 6 or 7 lines with DBH NA: 5 are dead trees; 1 is a crossed-out line on the data sheet
#fill in mean dbh for the one tree with dbh not recorded. Hot plot 3 50, RO.
#mean dbh of the other 6 red oaks in the plot is (14+20+14+16+12+14)6 = 15 inches.
#there is one tree (hotplot 3 16) with species = NA but sppgr=other hdwd. Keep.
treep$DBH<-ifelse(treep$hotspot==2 & treep$point==50 & treep$line==35, 15, treep$DBH)
treep<-subset(treep, !is.na(treep$DBH))

#get diameters into cm
treep$dbh.cm<-ifelse(!is.na(treep$DBH),treep$DBH*2.54, 0)
summary(treep)

#create species groups
levels(treep$SPP)
treep$sppgr<-ifelse(treep$SPP %in% c('BO','RO','SO','WO','CO','SWO'), 'oak', 
                    ifelse(treep$SPP %in% c('WP','HK','RP','NS','LA','RS'), 
                           'conifer','otherhd'))

#make a column with defoliation class represented as the mid-point proportion of that class
treep$DefProp<-ifelse(treep$RefClss==1, 0.125, ifelse(treep$RefClss==2, 0.375, ifelse(treep$RefClss==3, 0.625, 0.875)))

#make a column with a unique plot ID
treep$hotplot<-paste(treep$hotspot, treep$point)

#save this clean version of tree data
write.csv(treep, '2017HotSpotTrees.csv',row.names=FALSE)

## Summarize by Plot ##

## Some clean-up.
plotBAdef<-ddply(subset(treep, !is.na(treep$dbh.cm)), c('hotspot','point', 'hotplot' ,'type'),
               summarise,
               BAft2acre=length(point)*10,
               n.trees=length(point),
               MeanDef=mean(DefProp, na.rm=TRUE)) 

#iszero<-subset(treep, treep$dbh.cm==0) 
#3 lines have dbh=0. They represent plots with BA=0 (recent harvest)
#force the no-tree plots from 10 to 0 BA
plotBAdef$BAft2acre<-ifelse(plotBAdef$BAft2acre==10, 0, plotBAdef$BAft2acre) #yes, this pulls out only those 3 plots with no trees
plotBAdefno0<- subset(plotBAdef, !is.na(plotBAdef$MeanDef))

BApointspdef<-ddply(treep, c('hotspot','point', 'hotplot', 'sppgr'),
                   summarise,
                   BAft2acre=length(point)*10,
                   MeanDef=mean(DefProp, na.rm=TRUE)) # takes care of missing obs for one tree each in 1 7 and 3 4
BApointspdefno0<- subset(BApointspdef, !is.na(BApointspdef$MeanDef))

library(reshape2)
wspec <- reshape(BApointspdefno0, v.names = c("BAft2acre","MeanDef"), idvar = c('hotplot'), timevar = "sppgr", direction = "wide")
together<-cbind(plotBAdefno0, wspec)
together[is.na(together)] <- 0

#make sure this is correct
together$okay<-(together$BAft2acre.conifer+together$BAft2acre.oak+together$BAft2acre.otherhd)
together$test<-together$BAft2acre-together$okay
summary(together)
#add proportion oak and clean up the data frame
together$PropOak<-together$BAft2acre.oak/together$BAft2acre
HotPlotMeans<-together[, c(1:7, 11:16, 19)]

write.csv(HotPlotMeans, 'HotPlotMeans.csv', row.names = FALSE)
```