1.CleaningData_Bocas2021-2022_gh.Rmd

---
title: "Cleaning data Bocas 2021-2022"
output: html_document
date: '2024-07-22'
author: "Camila Calderon"
---

## load packages

```{r setup}
pacman::p_load(tidyverse,data.table, move, lubridate, cowplot, ggplot2, doBy, dplyr,  ggpubr, spatialrisk, rgdal, janitor, ISLR)
```

## Loading movebank data from 2021-2022

```{r data from movebank}
# download from movebank
me <- movebankLogin(username="", password="")
studyId <- getMovebankID("Greater spear-nosed bat (Phyllostomus hastatus) in Bocas del Toro 2021-2022",me)
batsmove<- getMovebankData(study=studyId, login=me)
batsdf <- as.data.frame(batsmove)
#unique(batsdf$tag_local_identifier)

```

## Changingm the names of some tag ids

```{r  editing data, results='hide'}
# change the ST of the tag local identifier of some individuals to TS
batsdf$tag_local_identifier <- gsub('ST', 'TS', batsdf$tag_local_identifier)
unique(batsdf$tag_local_identifier)
```
# Setting timestamp as posixct class

```{r remove outliers, include=TRUE}
batsdf$timestamp <- as.POSIXct(batsdf$timestamp, format = "%Y-%m-%d %H:%M:%S", tz ="UTC")

# to check if there is outliers
plot(batsdf$location_long, batsdf$location_lat)
```

## Removing some individuals with no good data

```{r}
# splitting data frame to a list
bocas.split <- split(batsdf, batsdf$tag_local_identifier)

# plot to see individuals
plot(bocas.split[["0C506E35_C"]]$location_long,bocas.split[["0C506E35_C"]]$location_lat)
plot(bocas.split[["5C500E4E_C"]]$location_long,bocas.split[["5C500E4E_C"]]$location_lat)
plot(bocas.split[["31507235_C"]]$location_long,bocas.split[["31507235_C"]]$location_lat)
plot(bocas.split[["53507135_C"]]$location_long,bocas.split[["53507135_C"]]$location_lat)
plot(bocas.split[["38506D37_C"]]$location_long,bocas.split[["38506D37_C"]]$location_lat)
plot(bocas.split[["38500337_C"]]$location_long,bocas.split[["38500337_C"]]$location_lat)
plot(bocas.split[["PH_TS_030"]]$location_long,bocas.split[["PH_TS_030"]]$location_lat)

# remove individuals where data is very incomplete 
bocas.splitnew <- bocas.split[names(bocas.split) %in% c("5C500E4E_C","PH_TS_030", "31507235_C", "53507135_C", "32501535_C")== FALSE] 

# create data frame
batsbocas.clean <- do.call(rbind, bocas.splitnew)

# split data frame
batsbocas.clean$ID <- paste0(batsbocas.clean$tag_local_identifier, "_", date(batsbocas.clean$timestamp))
allbats.batday <- split(batsbocas.clean, batsbocas.clean$ID)

# remove more individual days from march where data was not good 
batsclean <- allbats.batday[names(allbats.batday) %in%  c("0C506E35_C_2021-12-10", "0C506E35_C_2021-12-11", "22507B4E_D_2022-02-01", "2C500235_D_2022-02-13", "38506D37_D_2022-02-03", "0D501435_C_2021−12−09", "2D507235_G_2022-01-31", "37507537_D_2022-02-01",  "38506D37_C_2021-12-09",  "38506737_G_2022-01-30",  "39500E37_D_2022-02-01", "39500E37_D_2022-02-02", "PH_TS_016_2022-03-13", "PH_TS_016_2022-03-15", "PH_TS_018_2022-03-17", "PH_TS_018_2022-03-18", "PH_TS_018_2022-03-19", "PH_TS_024_2022-03-08", "PH_TS_024_2022-03-11", "PH_TS_024_2022-03-12", "PH_TS_024_2022-03-13", "PH_TS_024_2022-03-14", "PH_TS_024_2022-03-15", "PH_TS_039_2022-03-15", "PH_TS_049_2022-03-11", "PH_TS_049_2022-03-12", "PH_TS_049_2022-03-13", "PH_TS_052_2022-03-13", "PH_TS_052_2022-03-14", "PH_TS_056_2022-03-10", "PH_TS_056_2022-03-12", "PH_TS_056_2022-03-15", "PH_TS_056_2022-03-16", "PH_TS_056_2022-03-17", "PH_TS_062_2022-03-15", "PH_TS_062_2022-03-16", "PH_TS_062_2022-03-17", "PH_TS_072_2022-03-14", "PH_TS_072_2022-03-15", "PH_TS_072_2022-03-17", "PH_TS_072_2022-03-18", "PH_TS_074_2022-03-13", "PH_TS_074_2022-03-14", "PH_TS_081_2022-03-14", "PH_TS_081_2022-03-15", "PH_TS_081_2022-03-17", "PH_TS_083_2022-03-12", "PH_TS_083_2022-03-14", "PH_TS_083_2022-03-15", "PH_TS_085_2022-03-09", "PH_TS_085_2022-03-10", "PH_TS_085_2022-03-13", "PH_TS_085_2022-03-15", "PH_TS_085_2022-03-16", "PH_TS_096_2022-03-13", "PH_TS_096_2022-03-14", "PH_TS_096_2022-03-15", "PH_TS_098_2022-03-10", "PH_TS_098_2022-03-11", "PH_TS_100_2022-03-12", "PH_TS_100_2022-03-13", "PH_TS_100_2022-03-14", "PH_TS_100_2022-03-15", "PH_TS_103_2022-03-16", "PH_TS_113_2022-03-11", "PH_TS_113_2022-03-13", "PH_TS_113_2022-03-14", "PH_TS_113_2022-03-15", "PH_TS_114_2022-03-08", "PH_TS_114_2022-03-10", "PH_TS_121_2022-03-12","PH_TS_121_2022-03-13", "PH_TS_121_2022-03-15")== FALSE]

batsclean.df <- do.call(rbind, batsclean)
```

## Downsampling data from 2021 from the wet season

Now I am downsampling GPS data from December 2021 that was collected every 30 seconds. I will down sample to every 2 minutes to match the sampling rate from the data from February

```{r resampling data from december, results=FALSE}
# order by ID and timestamp
batsclean.df <- batsclean.df[order(batsclean.df$tag_local_identifier, batsclean.df$timestamp),]

# take data only from December
bats_wet <- batsclean.df %>%
  filter(date(timestamp)>"2021-12-01" & date(timestamp)<"2021-12-31")

# split by ID
bats_wet_lst<- split(bats_wet, bats_wet$ID)

library(amt)

# convert data frame to track object
resample <-  lapply(bats_wet_lst, function(x){
  mk_track(tbl=x, all_cols=T,
           .x=location_long, .y=location_lat, crs = 32617,
           .t=timestamp, order_by_ts = T, check_duplicates = T)
})

# resample every two minutes
resample_2 <- lapply(resample, function(x){
  track_resample(x, rate = minutes(2), tolerance = seconds(1), start = 1)
})

# check if resampling went well
lapply(resample_2, function(x){
  ggplot(aes(x=x_, y=y_), data=x) + geom_point() + ggtitle(x$ID)
})

# convert to dataframe again
batsdfwet <- do.call(rbind.data.frame,resample_2)
batsdfwet <- as.data.frame(batsdfwet)
names(batsdfwet)[1] <- "location_long"
names(batsdfwet)[2] <- "location_lat"
names(batsdfwet)[3] <- "timestamp"

# remove burst columns
batsdfwet <- batsdfwet[,-64]

# data frame without wet season
batsclean_nowet <- batsclean.df %>%
  filter(date(timestamp)>"2021-12-31")

# join to large data set of all bacas
bats_clean <- rbind(batsclean_nowet,batsdfwet)

# order data frame by id and timestamp
bats_clean[order(bats_clean$tag_local_identifier, bats_clean$timestamp),]

# it runs if the data is downloaded from movebank
movebats <- move(x=bats_clean$location_long, y=bats_clean$location_lat, 
	     time=as.POSIXct(bats_clean$timestamp, format="%Y-%m-%d %H:%M:%S", tz="UTC"), 
	     data=bats_clean, proj=CRS("+proj=longlat +ellps=WGS84"), 
	     animal=bats_clean$tag_local_identifier, sensor="GPS")

#"proj=utm +zone=17 +ellps=GRS80 +datum=NAD83 +units=m +no_defs"
n.indiv(movebats)

move::plot(movebats)

# initial movement parameters pre-cleaning.
moveList <- lapply(split(movebats), function(myInd){
  datechange <- c(0, abs(diff(as.numeric(as.factor(date(myInd@timestamps-(12*60*60)))))))
  myInd$BatDay <- cumsum(datechange)+1
  myInd$tlag <- c(NA,timeLag(myInd, units="secs"))
  myInd$step <- c(NA, move::distance(myInd))
  myInd$tag_ground_speed <- myInd$ground_speed
  myInd$ground_speed <- c(NA, move::speed(myInd))
  myInd$angle <- c(NA, angle(myInd))
  return(myInd)
})

movebats <- moveStack(moveList, forceTz="UTC")
movebats<- spTransform(movebats, CRS("+proj=utm +zone=17 +datum=WGS84"))
crds <- as.data.frame(movebats@coords)
movebats$x <- crds$coords.x1 
movebats$y <- crds$coords.x2 
# movebats <- spTransform(movebats, CRS("+proj=longlat +datum=WGS84"))
batsdf_new<- as.data.frame(movebats)

# create a column with id and batday
batsdf_new$ID_batday <- paste(batsdf_new$tag_local_identifier, batsdf_new$BatDay, sep="_")

#create column with cave id
batsdf_new$cave <- sapply(strsplit(batsdf_new$group_id,split = '-'), "[", 2)

#saving data
save(batsdf_new, file="~/ownCloud/PhDLife/P.hastatus/Thesis/paper1/BiologyLetters/data/BatsBocas_nooutliers.RData")

```