R/IRwolves.R
In tpetroel/PredClusters: Identification of Carnivore Predation Clusters from GPS Locations
Defines functions
IRwolves
Documented in
IRwolves
#' Isle Royale Wolf GPS Cluster Identification
#'
#' This function formats GPS data from Telonics and Vectoronic Collars used in Isle Royale National Park and then identifies potential predation sites ("clusters") from the formated GPS locations and user defined thresholds.
#' @param DownloadDate The date and folder title in which the raw GPS data are housed, should be formated as DDMONYYYY (i.e., 14MAY2019)
#' @param CollarID The individual identifier for each collar preceeded by "Collar" (i.e., Collar32712)
#' @param AnimalID The individual identifier for each wolf (i.e., W001F) indicating the wolf number and sex
#' @param NextCluster The next number to be assinged for cluster labeling for independent identification, reference Download_Dates.xlsx
#' @param ClusterDate The oldest date to be used for the period to identify clusters within (format = "YYYY-MM-DD", i.e., "2019-05-14")
#' @param time_thres Time, measured in days, between GPS locations to be considered as part of a cluster (i.e., 1 = 24 hours)
#' @param dist_thres Distance, measured in kilometers, between GPS points to be considered in a cluster (i.e., 0.05 = 50 meters)
#' @keywords Wolf Predation Clusters GPS Isle Royale
#' @examples
#' # Identify clusters formed by GPS locations that occur within 1 day of eachother
#' # and within 50 meters of eachother.
#'
#' IRwolves <- function(DownloadDate = "14MAY2019", CollarID = "Collar32712", AnimalID = "W001F", NextCluster = 001, ClusterDate = "2019-05-07", time_thres = 1, dist_thres = 0.05)
IRwolves = function(DownloadDate = DownloadDate, CollarID = CollarID, AnimalID = AnimalID,
NextCluster = NextCluster, ClusterDate = ClusterDate, time_thres = time_thres, dist_thres = dist_thres){
setwd(paste("C:/IR_Clusters/Downloads/",DownloadDate,sep=''))
# IF/ELSE STATEMENT FOR VECTRONIC VS. TELONICS ----------------------------
if (as.numeric(str_sub(CollarID,7))>678910){
###IF TELONICS
raw.data = data.frame(read.csv(paste("GPS_",CollarID,"_",AnimalID,".csv",sep=""),skip=23,header=T)) # Load the file name with the raw GPS data
str(raw.data)
raw.data$index <- as.numeric(row.names(raw.data))
raw.data <- raw.data[order(raw.data$index), ]
raw.data = subset(raw.data,GPS.Fix.Attempt=='Succeeded')
raw.data$line <- 1:nrow(raw.data)
time.adj = (data.frame(line=raw.data$line,UTC.Date=raw.data$GPS.Fix.Time,date=raw.data$GPS.Fix.Time,
UTC.Time=raw.data$GPS.Fix.Time,Time=raw.data$GPS.Fix.Time,lat=raw.data$GPS.Latitude,long=raw.data$GPS.Longitude,activity=1,
GPS.DT=raw.data$GPS.Fix.Time,ACT.DT=raw.data$GPS.Fix.Time))
time.adj$UTC.Date<-as.Date(time.adj$UTC.Date,format='%Y.%m.%d %H:%M:%S')
time.adj$GPS.DT<-strptime(time.adj$Time,format='%Y.%m.%d %H:%M:%S')
time.adj$ACT.DT<-strptime(time.adj$Time,format='%Y.%m.%d %H:%M:%S')
time.adj$UTC.Time<-strptime(time.adj$UTC.Time,format='%Y.%m.%d %H:%M:%S')
time.adj$UTC.Time<-strftime(time.adj$UTC.Time,format='%H:%M:%S')
time.adj$UTC.Time<-times(as.character(time.adj$UTC.Time))
time.adj$GPS.DT<-as.POSIXct(time.adj$GPS.DT,tz="Etc/UTC",usetz=TRUE)
time.adj$ACT.DT<-as.POSIXct(time.adj$ACT.DT,tz="Etc/UTC",usetz=TRUE)
# This formats date to work with the "find_cluster" function
time.adj$LMT.DT<-format(time.adj$GPS.DT,tz="Etc/GMT+5",usetz=TRUE)
# tz="Etc/GMT+5" will not take into account Daylight Savings, always keeps time -5 hours (EST) from GMT
# tz="America/Detroit" accounts for daylight savings (i.e., EST/EDT)
time.adj$Time<-strftime(time.adj$LMT.DT,format='%H:%M:%S')
time.adj$Time<-times(as.character(time.adj$Time))
time.adj$date<-as.Date(time.adj$LMT.DT,format='%Y-%m-%d %H:%M:%S')
clean.data = (data.frame(line=time.adj$line,Date=time.adj$UTC.Date,date=time.adj$date,
Time=time.adj$Time,lat=time.adj$lat,long=time.adj$long,activity=1,
GPS.DT=time.adj$GPS.DT,ACT.DT=time.adj$ACT.DT))
data = subset(clean.data,as.Date(date) >= ClusterDate & as.Date(date) < Sys.Date())
data$line <- as.numeric(row.names(data))
data <- data[order(data$line), ]
} else {
#####IF VECTRONIC
raw.data = data.frame(read.csv(paste('GPS_',CollarID,'_',AnimalID,'.csv',sep=""))) # Load the file name with the raw GPS data
#raw.act = data.frame(read.csv(paste("ACT_",CollarID,"_",DownloadDate,".csv",sep=""))) # Load the file name with the raw Activity data
#raw.act$sum.act = raw.act$ActivityX + raw.act$ActivityY # Sum X and Y Activity into one column
raw.data$GPS.dt = paste(raw.data$LMT_Date, raw.data$LMT_Time)
#raw.act$act.dt = paste(raw.act$LMT_Date, raw.act$LMT_Time)
raw.data$GPS.dt<-as.POSIXct(raw.data$GPS.dt,format="%m/%d/%Y %H:%M:%S")
#raw.act$act.dt<-as.POSIXct(raw.act$act.dt,format="%m/%d/%Y %H:%M:%S")
#raw.data$GPS.dt = as.POSIXct(paste(raw.data$LMT_Date, raw.data$LMT_Time),format="%Y-%m-%d %H:%M:%S")
#raw.act$act.dt = as.POSIXct(paste(raw.act$LMT_Date, raw.act$LMT_Time),format="%Y-%m-%d %H:%M:%S")
#raw.merge = cbind(raw.data, raw.act[ sapply(raw.data$GPS.dt, function(x) which.min(abs(difftime(x, raw.act$act.dt)))), ])
clean.data = na.omit(data.frame(line=raw.data$No,Date=raw.data$LMT_Date,date=raw.data$LMT_Date,
Time=raw.data$LMT_Time,lat=raw.data$Latitude,long=raw.data$Longitude,activity=1,
GPS.DT=raw.data$GPS.dt,ACT.DT=raw.data$GPS.dt))
# Cleans up missing data only keeps specified columns
clean.data$Date = as.character(clean.data$Date)
clean.data$Date = as.Date(clean.data$Date,format="%m/%d/%Y")
clean.data$date = as.character(clean.data$date)
clean.data$date = as.Date(clean.data$date,format="%m/%d/%Y")
# This formats date to work with the "find_cluster" function
data = subset(clean.data,as.Date(date) >= ClusterDate & as.Date(date) < Sys.Date())
}
# EXPORT PRIMARY GPS FILE ------------------------------
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,sep=''))
data$AnimalID = AnimalID
write.csv(data,paste(AnimalID,CollarID,DownloadDate,'R.csv',sep="_"))
data$PID = 1
data$POS = data$line
## function 'as.PolySet' builds PolySet-class from data.frame for 'convUL' to convert latlong to UTM
data.XY = data.frame(POS=data$POS,PID=data$PID,X=data$long,Y=data$lat) # "X"=longitude "Y"=latitude in PBSmapping
data.LL = as.PolySet(data.XY, projection = "LL", zone = 16) # Change LongLat to Poly Set
data.UTM = convUL(data.LL,km = FALSE) # Convert LongLat to UTM
data.locs = data.UTM[, c("X", "Y")] # Extract "Easting"& "Northing"
data = cbind(data,data.locs)
data = subset(data, select = -c(GPS.DT, ACT.DT))
setwd(paste("C:/IR_Clusters/Downloads/",DownloadDate,sep=""))
write.table(data,paste("ALL_WOLF_GPS_DATA",DownloadDate,".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=T) # Output of cluster locations/date/time
b<-read.table(paste("ALL_WOLF_GPS_DATA",DownloadDate,".txt",sep = ""),fill=T,na.strings="NA",header=T)
b<-b[!b$AnimalID =="AnimalID",]
b$X <-as.character(b$X)
b$Y <-as.character(b$Y)
b$X <-as.numeric(b$X)
b$Y <-as.numeric(b$Y)
b$id = 1:nrow(b)
write.table(b,paste("ALL_WOLF_GPS_DATA",DownloadDate,".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=F) # Output of cluster locations/date/time
ST = data.frame(Id = b$id, Easting = b$X, Northing = b$Y)
AT = data.frame(AnimalID = b$AnimalID, Line = b$line, Date = b$Date,
Time = b$Time, Id = b$id)
class(ST$Id)
class(AT$Id)
shpfl = convert.to.shapefile(shpTable = ST, attTable = AT, field = "Id",type = 1)
class(shpfl)
write.shapefile(shpfl, paste("ALL_WOLF_GPS_DATA",DownloadDate,sep = ""),arcgis = TRUE) # Output of shape file for ArcGIS mapping of cluster locations
## RUN FUNCTION WITH DEFINED PARAMETERS ----------------------------------------------------------------------------------------------------------
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,'/Tables_Maps',sep=''))
## Function command line, "time_thres" is days between first/last point (i.e., 1=24hr),
# "dist_thres" is km boundry to include points (0.05 = 50m),
# "size" refers to text size on cluster map
res1 = find_cluster(ind = AnimalID, data = data, time_thres = time_thres, dist_thres = dist_thres, size = 2)
res = read.table("res.txt" , header=T, sep="\t", dec=".")
centr = read.table("centr.txt", header=T, sep="\t", dec=".")
act.clus = subset(res, cluster >0)
act.clus = aggregate(activity ~ cluster, data=act.clus,FUN=sum)
res$PID = 1
centr$PID = 1
res$POS = res$line
centr$POS = centr$cluster
## function 'as.PolySet' builds PolySet-class from data.frame for 'convUL' to convert latlong to UTM
res.XY = data.frame(POS=res$POS,PID=res$PID,X=res$long,Y=res$lat) # "X"=longitude "Y"=latitude in PBSmapping
centr.XY = data.frame(POS=centr$POS,PID=centr$PID,X=centr$long,Y=centr$lat)
res.LL = as.PolySet(res.XY, projection = "LL", zone = 16) # Change LongLat to Poly Set
res.UTM = convUL(res.LL,km = FALSE) # Convert LongLat to UTM
res.locs = res.UTM[, c("X", "Y")] # Extract "Easting"& "Northing"
res = cbind(res,res.locs)
centr.LL = as.PolySet(centr.XY, projection = "LL", zone = 16) # Change LongLat to Poly Set
centr.UTM = convUL(centr.LL,km = FALSE) # Convert LongLat to UTM
centr.locs = centr.UTM[, c("X", "Y")] # Extract "Easting"& "Northing"
clus.centr = cbind(centr,centr.locs,act.clus)
clus.centr$avg.act = round(clus.centr$activity/clus.centr$n, digits = 1)
clus.centrs = subset(clus.centr, n >= 2) # Takes a subset of all defined clusters with 5 or more locations
centrs = data.frame(cluster_ID=clus.centrs$cluster, n=clus.centrs$n, activity=clus.centrs$activity,
avg_act=clus.centrs$avg.act, Date_i=clus.centrs$Date_i, Time_i=clus.centrs$Time_i,
Date_f=clus.centrs$Date_f, Time_f=clus.centrs$Time_f, Easting=clus.centrs$X,Northing=clus.centrs$Y)
centrs[c("Initials","Date_Completed")] = ""
centrs[c("Animal_ID")] = AnimalID
centrs[c("Cluster_Num")] = seq(from = as.numeric(NextCluster), to = nrow(centrs)+ as.numeric(NextCluster) -1, by = 1)
centrs[c("Correct_Num")] = sprintf("%03d", centrs$Cluster_Num)
centrs[c("Cluster_ID")] = paste(centrs$Animal_ID, centrs$Correct_Num, sep = "_")
## This adds columns used for cluster tables. Initials and Date Completed are filled out when visiting a cluster.
## Animal_ID is changed each time for the corresponding animal for this collar data
## Cluster_Num is generating a sequence of numbers starting with the specified cluster number from the Download Dates file. This should be changed each time it is run.
## The from number will be the first cluster in the sequence
## The to number is the last number in this cluster sequence, calculate this by looking in the Global Environment at "centrs". The number of obs. is the number of clusters generated with this download.
## For example from= 125, to = 247, by 1, means there are 123 obs starting at cluster 125, going to 247, by increments of 1
## Correct_Num is correcting the number of digits, i.e. adding zeros, in the cluster number
## Cluster_ID is combining the column with Animal_ID and Corrected_Num to get the correct Cluster_ID for each cluster generated
centrs_clean = data.frame(Animal_ID=centrs$Animal_ID, Cluster_ID=centrs$Cluster_ID, n=clus.centrs$n, activity=clus.centrs$activity, avg_act=clus.centrs$avg.act, Date_i=centrs$Date_i, Time_i=centrs$Time_i,
Date_f=centrs$Date_f, Time_f=centrs$Time_f, Easting=clus.centrs$X,Northing=clus.centrs$Y, Initials=centrs$Initials, Date_Completed=centrs$Date_Completed)
## This creates a new data frame, "centrs_clean", with only the columns needed for cluster tables that is used below to write into a text file
centrs_clean$Easting <- round(centrs$Easting,0)
centrs_clean$Northing <- round(centrs$Northing,0)
## CLUSTER KILL SITE PREDICTION-----------------------------------------------------------------------------------------------------------
# Load fitted models for each carnivore species (BB.fm, BC.fm, CO.fm, WO.fm)
## Scale cluster covariates to match data from fitted models
#newdat <- as.data.frame((centrs_clean$n - 22.2748815) / 22.6240699)
#colnames(newdat) <- c("LOCS")
#newdat$MEANACT <- (centrs_clean$avg_act - 23.8128815) / 28.5886211
#newdat$SUMACT <- (centrs_clean$activity - 381.2426540) / 237.5869941
## Predict which clusters are likely kill sites from models built with Phase 2 data
## Does not apply to IR wolves, disregard errors here
#pred <- data.frame(predict(model, newdata=newdat, ## CHANGE MODEL HERE ACCORDING TO SPECIES USED (BB.fm, BC.fm, CO.fm, WO.fm)
# type = "response"))
#colnames(pred) <- c("logodds") # Prediction for kill sites in logodds
#pred$prob <- exp(pred[,1])/(1+exp(pred[,1])) # Covert the cut-of value out of logodds
## Cut-off Values --
#BBX <- 0.5051975
#BCX <- 0.5462662
#COX <- 0.5172079
#WOX <- 0.5269982
#for(i in 1:nrow(pred))
# if(pred$prob[i] > paste(substring(AnimalID,1,2),"X",sep="")) { ## CHANGE TO THE CORRECT SPECIES CUTOFF
# pred$Predicted[i] = "YES" # yes, the cluster is predicted as a kill
# } else {
# pred$Predicted[i] = "NO" # no, the cluster is likely not a kill
# }
## OUTPUT OF CLUSTERS TABLES/MAPS/ETC TO FILES---------------------------------------------------------------------------------------------------------
drops<-c("activity","avg_act")
centrs_clean <- centrs_clean[,!names(centrs_clean) %in% drops]
write.table(res,paste(AnimalID,"_",DownloadDate,"_res",".txt",sep = ""), sep="\t", dec=".",quote=FALSE,row.name=FALSE) # Output of all calculated clusters from function "find_Cluster"
write.table(centrs_clean,paste(AnimalID,"_",DownloadDate,"_center",".txt",sep = ""), sep="\t", dec=".",quote=FALSE,row.name=FALSE) # Output of cluster locations/date/time
setwd(paste("C:/IR_Clusters/Downloads/",DownloadDate,sep=""))
write.table(centrs_clean,paste("ALL_CLUSTERS",DownloadDate,"_center",".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=T) # Output of cluster locations/date/time
a<-read.table(paste("ALL_CLUSTERS",DownloadDate,"_center",".txt",sep = ""),fill=T,na.strings="NA",header=T)
a<-a[!a$Animal_ID=="Animal_ID",]
a$Easting<-as.character(a$Easting)
a$Northing<-as.character(a$Northing)
a$Easting<-as.numeric(a$Easting)
a$Northing<-as.numeric(a$Northing)
write.table(a,paste("ALL_CLUSTERS",DownloadDate,"_center",".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=F) # Output of cluster locations/date/time
shpTable = data.frame(Id = a$Cluster_ID, Easting = a$Easting, Northing = a$Northing)
attTable = data.frame(Animal = a$Animal_ID, Id = a$Cluster_ID, N = a$n, Date_Started = a$Date_i,
Time_Started = a$Time_i, Date_Finished = a$Date_f,
Time_Finished = a$Time_f, Cluster_ID = a$Cluster_ID)
class(shpTable$Id)
class(attTable$Id)
shp.file = convert.to.shapefile(shpTable = shpTable, attTable = attTable, field = "Id",type = 1)
class(shp.file)
write.shapefile(shp.file, paste("ALL_CLUSTERS",DownloadDate,"_center",sep = ""),arcgis = TRUE) # Output of shape file for ArcGIS mapping of cluster locations
## -------------------------------------------
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,'/Shapefiles',sep=''))
shpTable = data.frame(Id = centrs$cluster_ID, Easting = centrs$Easting, Northing = centrs$Northing)
attTable = data.frame(Id = centrs$cluster_ID, N = centrs$n, Date_Started = centrs$Date_i,
Time_Started = centrs$Time_i, Date_Finished = centrs$Date_f,
Time_Finished = centrs$Time_f, Cluster_ID = centrs_clean$Cluster_ID)
shp.file = convert.to.shapefile(shpTable = shpTable, attTable = attTable, field = "Id",type = 1)
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,"/Shapefiles",sep=''))
write.shapefile(shp.file, paste(AnimalID,"_",DownloadDate,"_center",sep = ""), arcgis = TRUE) # Output of shape file for ArcGIS mapping of cluster locations
}
tpetroel/PredClusters documentation built on Feb. 5, 2020, 5:49 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions IRwolves
Documented in IRwolves
#' Isle Royale Wolf GPS Cluster Identification
#'
#' This function formats GPS data from Telonics and Vectoronic Collars used in Isle Royale National Park and then identifies potential predation sites ("clusters") from the formated GPS locations and user defined thresholds.
#' @param DownloadDate The date and folder title in which the raw GPS data are housed, should be formated as DDMONYYYY (i.e., 14MAY2019)
#' @param CollarID The individual identifier for each collar preceeded by "Collar" (i.e., Collar32712)
#' @param AnimalID The individual identifier for each wolf (i.e., W001F) indicating the wolf number and sex
#' @param NextCluster The next number to be assinged for cluster labeling for independent identification, reference Download_Dates.xlsx
#' @param ClusterDate The oldest date to be used for the period to identify clusters within (format = "YYYY-MM-DD", i.e., "2019-05-14")
#' @param time_thres Time, measured in days, between GPS locations to be considered as part of a cluster (i.e., 1 = 24 hours)
#' @param dist_thres Distance, measured in kilometers, between GPS points to be considered in a cluster (i.e., 0.05 = 50 meters)
#' @keywords Wolf Predation Clusters GPS Isle Royale
#' @examples
#' # Identify clusters formed by GPS locations that occur within 1 day of eachother
#' # and within 50 meters of eachother.
#'
#' IRwolves <- function(DownloadDate = "14MAY2019", CollarID = "Collar32712", AnimalID = "W001F", NextCluster = 001, ClusterDate = "2019-05-07", time_thres = 1, dist_thres = 0.05)
IRwolves = function(DownloadDate = DownloadDate, CollarID = CollarID, AnimalID = AnimalID,
NextCluster = NextCluster, ClusterDate = ClusterDate, time_thres = time_thres, dist_thres = dist_thres){
setwd(paste("C:/IR_Clusters/Downloads/",DownloadDate,sep=''))
# IF/ELSE STATEMENT FOR VECTRONIC VS. TELONICS ----------------------------
if (as.numeric(str_sub(CollarID,7))>678910){
###IF TELONICS
raw.data = data.frame(read.csv(paste("GPS_",CollarID,"_",AnimalID,".csv",sep=""),skip=23,header=T)) # Load the file name with the raw GPS data
str(raw.data)
raw.data$index <- as.numeric(row.names(raw.data))
raw.data <- raw.data[order(raw.data$index), ]
raw.data = subset(raw.data,GPS.Fix.Attempt=='Succeeded')
raw.data$line <- 1:nrow(raw.data)
time.adj = (data.frame(line=raw.data$line,UTC.Date=raw.data$GPS.Fix.Time,date=raw.data$GPS.Fix.Time,
UTC.Time=raw.data$GPS.Fix.Time,Time=raw.data$GPS.Fix.Time,lat=raw.data$GPS.Latitude,long=raw.data$GPS.Longitude,activity=1,
GPS.DT=raw.data$GPS.Fix.Time,ACT.DT=raw.data$GPS.Fix.Time))
time.adj$UTC.Date<-as.Date(time.adj$UTC.Date,format='%Y.%m.%d %H:%M:%S')
time.adj$GPS.DT<-strptime(time.adj$Time,format='%Y.%m.%d %H:%M:%S')
time.adj$ACT.DT<-strptime(time.adj$Time,format='%Y.%m.%d %H:%M:%S')
time.adj$UTC.Time<-strptime(time.adj$UTC.Time,format='%Y.%m.%d %H:%M:%S')
time.adj$UTC.Time<-strftime(time.adj$UTC.Time,format='%H:%M:%S')
time.adj$UTC.Time<-times(as.character(time.adj$UTC.Time))
time.adj$GPS.DT<-as.POSIXct(time.adj$GPS.DT,tz="Etc/UTC",usetz=TRUE)
time.adj$ACT.DT<-as.POSIXct(time.adj$ACT.DT,tz="Etc/UTC",usetz=TRUE)
# This formats date to work with the "find_cluster" function
time.adj$LMT.DT<-format(time.adj$GPS.DT,tz="Etc/GMT+5",usetz=TRUE)
# tz="Etc/GMT+5" will not take into account Daylight Savings, always keeps time -5 hours (EST) from GMT
# tz="America/Detroit" accounts for daylight savings (i.e., EST/EDT)
time.adj$Time<-strftime(time.adj$LMT.DT,format='%H:%M:%S')
time.adj$Time<-times(as.character(time.adj$Time))
time.adj$date<-as.Date(time.adj$LMT.DT,format='%Y-%m-%d %H:%M:%S')
clean.data = (data.frame(line=time.adj$line,Date=time.adj$UTC.Date,date=time.adj$date,
Time=time.adj$Time,lat=time.adj$lat,long=time.adj$long,activity=1,
GPS.DT=time.adj$GPS.DT,ACT.DT=time.adj$ACT.DT))
data = subset(clean.data,as.Date(date) >= ClusterDate & as.Date(date) < Sys.Date())
data$line <- as.numeric(row.names(data))
data <- data[order(data$line), ]
} else {
#####IF VECTRONIC
raw.data = data.frame(read.csv(paste('GPS_',CollarID,'_',AnimalID,'.csv',sep=""))) # Load the file name with the raw GPS data
#raw.act = data.frame(read.csv(paste("ACT_",CollarID,"_",DownloadDate,".csv",sep=""))) # Load the file name with the raw Activity data
#raw.act$sum.act = raw.act$ActivityX + raw.act$ActivityY # Sum X and Y Activity into one column
raw.data$GPS.dt = paste(raw.data$LMT_Date, raw.data$LMT_Time)
#raw.act$act.dt = paste(raw.act$LMT_Date, raw.act$LMT_Time)
raw.data$GPS.dt<-as.POSIXct(raw.data$GPS.dt,format="%m/%d/%Y %H:%M:%S")
#raw.act$act.dt<-as.POSIXct(raw.act$act.dt,format="%m/%d/%Y %H:%M:%S")
#raw.data$GPS.dt = as.POSIXct(paste(raw.data$LMT_Date, raw.data$LMT_Time),format="%Y-%m-%d %H:%M:%S")
#raw.act$act.dt = as.POSIXct(paste(raw.act$LMT_Date, raw.act$LMT_Time),format="%Y-%m-%d %H:%M:%S")
#raw.merge = cbind(raw.data, raw.act[ sapply(raw.data$GPS.dt, function(x) which.min(abs(difftime(x, raw.act$act.dt)))), ])
clean.data = na.omit(data.frame(line=raw.data$No,Date=raw.data$LMT_Date,date=raw.data$LMT_Date,
Time=raw.data$LMT_Time,lat=raw.data$Latitude,long=raw.data$Longitude,activity=1,
GPS.DT=raw.data$GPS.dt,ACT.DT=raw.data$GPS.dt))
# Cleans up missing data only keeps specified columns
clean.data$Date = as.character(clean.data$Date)
clean.data$Date = as.Date(clean.data$Date,format="%m/%d/%Y")
clean.data$date = as.character(clean.data$date)
clean.data$date = as.Date(clean.data$date,format="%m/%d/%Y")
# This formats date to work with the "find_cluster" function
data = subset(clean.data,as.Date(date) >= ClusterDate & as.Date(date) < Sys.Date())
}
# EXPORT PRIMARY GPS FILE ------------------------------
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,sep=''))
data$AnimalID = AnimalID
write.csv(data,paste(AnimalID,CollarID,DownloadDate,'R.csv',sep="_"))
data$PID = 1
data$POS = data$line
## function 'as.PolySet' builds PolySet-class from data.frame for 'convUL' to convert latlong to UTM
data.XY = data.frame(POS=data$POS,PID=data$PID,X=data$long,Y=data$lat) # "X"=longitude "Y"=latitude in PBSmapping
data.LL = as.PolySet(data.XY, projection = "LL", zone = 16) # Change LongLat to Poly Set
data.UTM = convUL(data.LL,km = FALSE) # Convert LongLat to UTM
data.locs = data.UTM[, c("X", "Y")] # Extract "Easting"& "Northing"
data = cbind(data,data.locs)
data = subset(data, select = -c(GPS.DT, ACT.DT))
setwd(paste("C:/IR_Clusters/Downloads/",DownloadDate,sep=""))
write.table(data,paste("ALL_WOLF_GPS_DATA",DownloadDate,".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=T) # Output of cluster locations/date/time
b<-read.table(paste("ALL_WOLF_GPS_DATA",DownloadDate,".txt",sep = ""),fill=T,na.strings="NA",header=T)
b<-b[!b$AnimalID =="AnimalID",]
b$X <-as.character(b$X)
b$Y <-as.character(b$Y)
b$X <-as.numeric(b$X)
b$Y <-as.numeric(b$Y)
b$id = 1:nrow(b)
write.table(b,paste("ALL_WOLF_GPS_DATA",DownloadDate,".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=F) # Output of cluster locations/date/time
ST = data.frame(Id = b$id, Easting = b$X, Northing = b$Y)
AT = data.frame(AnimalID = b$AnimalID, Line = b$line, Date = b$Date,
Time = b$Time, Id = b$id)
class(ST$Id)
class(AT$Id)
shpfl = convert.to.shapefile(shpTable = ST, attTable = AT, field = "Id",type = 1)
class(shpfl)
write.shapefile(shpfl, paste("ALL_WOLF_GPS_DATA",DownloadDate,sep = ""),arcgis = TRUE) # Output of shape file for ArcGIS mapping of cluster locations
## RUN FUNCTION WITH DEFINED PARAMETERS ----------------------------------------------------------------------------------------------------------
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,'/Tables_Maps',sep=''))
## Function command line, "time_thres" is days between first/last point (i.e., 1=24hr),
# "dist_thres" is km boundry to include points (0.05 = 50m),
# "size" refers to text size on cluster map
res1 = find_cluster(ind = AnimalID, data = data, time_thres = time_thres, dist_thres = dist_thres, size = 2)
res = read.table("res.txt" , header=T, sep="\t", dec=".")
centr = read.table("centr.txt", header=T, sep="\t", dec=".")
act.clus = subset(res, cluster >0)
act.clus = aggregate(activity ~ cluster, data=act.clus,FUN=sum)
res$PID = 1
centr$PID = 1
res$POS = res$line
centr$POS = centr$cluster
## function 'as.PolySet' builds PolySet-class from data.frame for 'convUL' to convert latlong to UTM
res.XY = data.frame(POS=res$POS,PID=res$PID,X=res$long,Y=res$lat) # "X"=longitude "Y"=latitude in PBSmapping
centr.XY = data.frame(POS=centr$POS,PID=centr$PID,X=centr$long,Y=centr$lat)
res.LL = as.PolySet(res.XY, projection = "LL", zone = 16) # Change LongLat to Poly Set
res.UTM = convUL(res.LL,km = FALSE) # Convert LongLat to UTM
res.locs = res.UTM[, c("X", "Y")] # Extract "Easting"& "Northing"
res = cbind(res,res.locs)
centr.LL = as.PolySet(centr.XY, projection = "LL", zone = 16) # Change LongLat to Poly Set
centr.UTM = convUL(centr.LL,km = FALSE) # Convert LongLat to UTM
centr.locs = centr.UTM[, c("X", "Y")] # Extract "Easting"& "Northing"
clus.centr = cbind(centr,centr.locs,act.clus)
clus.centr$avg.act = round(clus.centr$activity/clus.centr$n, digits = 1)
clus.centrs = subset(clus.centr, n >= 2) # Takes a subset of all defined clusters with 5 or more locations
centrs = data.frame(cluster_ID=clus.centrs$cluster, n=clus.centrs$n, activity=clus.centrs$activity,
avg_act=clus.centrs$avg.act, Date_i=clus.centrs$Date_i, Time_i=clus.centrs$Time_i,
Date_f=clus.centrs$Date_f, Time_f=clus.centrs$Time_f, Easting=clus.centrs$X,Northing=clus.centrs$Y)
centrs[c("Initials","Date_Completed")] = ""
centrs[c("Animal_ID")] = AnimalID
centrs[c("Cluster_Num")] = seq(from = as.numeric(NextCluster), to = nrow(centrs)+ as.numeric(NextCluster) -1, by = 1)
centrs[c("Correct_Num")] = sprintf("%03d", centrs$Cluster_Num)
centrs[c("Cluster_ID")] = paste(centrs$Animal_ID, centrs$Correct_Num, sep = "_")
## This adds columns used for cluster tables. Initials and Date Completed are filled out when visiting a cluster.
## Animal_ID is changed each time for the corresponding animal for this collar data
## Cluster_Num is generating a sequence of numbers starting with the specified cluster number from the Download Dates file. This should be changed each time it is run.
## The from number will be the first cluster in the sequence
## The to number is the last number in this cluster sequence, calculate this by looking in the Global Environment at "centrs". The number of obs. is the number of clusters generated with this download.
## For example from= 125, to = 247, by 1, means there are 123 obs starting at cluster 125, going to 247, by increments of 1
## Correct_Num is correcting the number of digits, i.e. adding zeros, in the cluster number
## Cluster_ID is combining the column with Animal_ID and Corrected_Num to get the correct Cluster_ID for each cluster generated
centrs_clean = data.frame(Animal_ID=centrs$Animal_ID, Cluster_ID=centrs$Cluster_ID, n=clus.centrs$n, activity=clus.centrs$activity, avg_act=clus.centrs$avg.act, Date_i=centrs$Date_i, Time_i=centrs$Time_i,
Date_f=centrs$Date_f, Time_f=centrs$Time_f, Easting=clus.centrs$X,Northing=clus.centrs$Y, Initials=centrs$Initials, Date_Completed=centrs$Date_Completed)
## This creates a new data frame, "centrs_clean", with only the columns needed for cluster tables that is used below to write into a text file
centrs_clean$Easting <- round(centrs$Easting,0)
centrs_clean$Northing <- round(centrs$Northing,0)
## CLUSTER KILL SITE PREDICTION-----------------------------------------------------------------------------------------------------------
# Load fitted models for each carnivore species (BB.fm, BC.fm, CO.fm, WO.fm)
## Scale cluster covariates to match data from fitted models
#newdat <- as.data.frame((centrs_clean$n - 22.2748815) / 22.6240699)
#colnames(newdat) <- c("LOCS")
#newdat$MEANACT <- (centrs_clean$avg_act - 23.8128815) / 28.5886211
#newdat$SUMACT <- (centrs_clean$activity - 381.2426540) / 237.5869941
## Predict which clusters are likely kill sites from models built with Phase 2 data
## Does not apply to IR wolves, disregard errors here
#pred <- data.frame(predict(model, newdata=newdat, ## CHANGE MODEL HERE ACCORDING TO SPECIES USED (BB.fm, BC.fm, CO.fm, WO.fm)
# type = "response"))
#colnames(pred) <- c("logodds") # Prediction for kill sites in logodds
#pred$prob <- exp(pred[,1])/(1+exp(pred[,1])) # Covert the cut-of value out of logodds
## Cut-off Values --
#BBX <- 0.5051975
#BCX <- 0.5462662
#COX <- 0.5172079
#WOX <- 0.5269982
#for(i in 1:nrow(pred))
# if(pred$prob[i] > paste(substring(AnimalID,1,2),"X",sep="")) { ## CHANGE TO THE CORRECT SPECIES CUTOFF
# pred$Predicted[i] = "YES" # yes, the cluster is predicted as a kill
# } else {
# pred$Predicted[i] = "NO" # no, the cluster is likely not a kill
# }
## OUTPUT OF CLUSTERS TABLES/MAPS/ETC TO FILES---------------------------------------------------------------------------------------------------------
drops<-c("activity","avg_act")
centrs_clean <- centrs_clean[,!names(centrs_clean) %in% drops]
write.table(res,paste(AnimalID,"_",DownloadDate,"_res",".txt",sep = ""), sep="\t", dec=".",quote=FALSE,row.name=FALSE) # Output of all calculated clusters from function "find_Cluster"
write.table(centrs_clean,paste(AnimalID,"_",DownloadDate,"_center",".txt",sep = ""), sep="\t", dec=".",quote=FALSE,row.name=FALSE) # Output of cluster locations/date/time
setwd(paste("C:/IR_Clusters/Downloads/",DownloadDate,sep=""))
write.table(centrs_clean,paste("ALL_CLUSTERS",DownloadDate,"_center",".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=T) # Output of cluster locations/date/time
a<-read.table(paste("ALL_CLUSTERS",DownloadDate,"_center",".txt",sep = ""),fill=T,na.strings="NA",header=T)
a<-a[!a$Animal_ID=="Animal_ID",]
a$Easting<-as.character(a$Easting)
a$Northing<-as.character(a$Northing)
a$Easting<-as.numeric(a$Easting)
a$Northing<-as.numeric(a$Northing)
write.table(a,paste("ALL_CLUSTERS",DownloadDate,"_center",".txt",sep = ""),sep="\t", dec=".",quote=FALSE,row.name=FALSE,append=F) # Output of cluster locations/date/time
shpTable = data.frame(Id = a$Cluster_ID, Easting = a$Easting, Northing = a$Northing)
attTable = data.frame(Animal = a$Animal_ID, Id = a$Cluster_ID, N = a$n, Date_Started = a$Date_i,
Time_Started = a$Time_i, Date_Finished = a$Date_f,
Time_Finished = a$Time_f, Cluster_ID = a$Cluster_ID)
class(shpTable$Id)
class(attTable$Id)
shp.file = convert.to.shapefile(shpTable = shpTable, attTable = attTable, field = "Id",type = 1)
class(shp.file)
write.shapefile(shp.file, paste("ALL_CLUSTERS",DownloadDate,"_center",sep = ""),arcgis = TRUE) # Output of shape file for ArcGIS mapping of cluster locations
## -------------------------------------------
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,'/Shapefiles',sep=''))
shpTable = data.frame(Id = centrs$cluster_ID, Easting = centrs$Easting, Northing = centrs$Northing)
attTable = data.frame(Id = centrs$cluster_ID, N = centrs$n, Date_Started = centrs$Date_i,
Time_Started = centrs$Time_i, Date_Finished = centrs$Date_f,
Time_Finished = centrs$Time_f, Cluster_ID = centrs_clean$Cluster_ID)
shp.file = convert.to.shapefile(shpTable = shpTable, attTable = attTable, field = "Id",type = 1)
setwd(paste("C:/IR_Clusters/Wolf_GPS_Data/",AnimalID,"/Shapefiles",sep=''))
write.shapefile(shp.file, paste(AnimalID,"_",DownloadDate,"_center",sep = ""), arcgis = TRUE) # Output of shape file for ArcGIS mapping of cluster locations
}
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.