2.CleaningData_2023_gh.Rmd

---
title: "Cleaning 2023 data"
author: "Camila Calderon"
date: "2024-07-22"
output: html_document
---

## Load libraries 

```{r, warning=FALSE, message = FALSE}
library(move)
library(data.table)
library(janitor)
library(lubridate)
library(amt)
library(ggplot2)
```

## Download data from movebank for Bocas 2023

```{r}
me <- movebankLogin(username="", password="")
studyId <- getMovebankID("Greater spear-nosed bats (Phyllostomus hastatus) in Bocas del Toro 2023",me)
batsBocas<- getMovebankData(study=studyId, login=me)
n.indiv(batsBocas)

# plot tracks
move::plot(batsBocas)
```

## Removing data which is incomplete or not good

```{r pressure, results=FALSE, fig.keep = "first"}
# convert move object to data frame for cleaning
batsBocas_df <- as.data.frame(batsBocas)

# add id_day column
batsBocas_df$ID <- paste(batsBocas_df$tag_local_identifier, date(batsBocas_df$timestamp), sep="_")

# plotting individuals to see which ones have good data by ID
lapply(split(batsBocas_df, batsBocas_df$ID), function(x){
  ggplot(x, aes(x=timestamp, y=location_long))+geom_point()+ggtitle(x$ID)
})

# plot by day
# lapply(split(batsBocas_df, date(batsBocas_df$timestamp)), function(x){
#     ggplot(aes(x=location_long, y=location_lat, color=tag_local_identifier), data=x)+
#     geom_point()+
#     ggtitle(date(x$timestamp))
# })
```

## Remove days with bad data

```{r}
# split data frame by id
batsBocas_ls <- split(batsBocas_df, batsBocas_df$ID)

batsBocas_clean_ls <- batsBocas_ls[names(batsBocas_ls) %in% c("PHYL4_2023-08-12","PHYL4_2023-08-17", "PHYL40_2023-08-12", "PHYL40_2023-08-16","PHYL40_2023-08-17", "PHYL7_2023-08-14", "PHYL38_2023-08-16", "PHYL36_2023-08-12", "PHYL36_2023-08-12", "PHYL36_2023-08-13", "PHYL36_2023-08-15", "PHYL36_2023-08-16", "PHYL36_2023-08-17", "PHYL34_2023-08-12", "PHYL34_2023-08-15", "PHYL34_2023-08-16","PHYL33_2023-08-12","PHYL33_2023-08-14", "PHYL33_2023-08-16", "PHYL32_2023-08-12", "PHYL32_2023-08-14", "PHYL31_2023-08-16","PHYL30_2023-08-17","PHYL3_2023-08-15","PHYL29_2023-08-13", "PHYL24_2023-08-18", "PHYL22_2023-08-16", "PHYL22_2023-08-17", "PHYL21_2023-08-15") == FALSE] 
# bat 37 is not great but I still will leave it 
# bats 22, 24, 32, 35, 39, 30 forage in the island at least one of the days.
# bat 22 forage on the island on August 14 and 15
# bat 24 forage on the island on August 14-18

# plot data of the individuals to be used
lapply(batsBocas_clean_ls, function(x){
    ggplot(aes(x=location_long, y=location_lat), data=x)+
    geom_point()+
    ggtitle(unique(x$ID))
})
```
## Renaming group and sex information to the data

```{r, results=FALSE}
# covert list of ids to data frame again
batsBocas_cleandf <- do.call(rbind, batsBocas_clean_ls)

# order by ID and timestamp
batsBocas_cleandf <- batsBocas_cleandf[order(batsBocas_cleandf$tag_local_identifier, batsBocas_cleandf$timestamp),]

# add cave group
batsBocas_cleandf$cave_group <-batsBocas_cleandf$group_id

# add cave only 
batsBocas_cleandf$cave <- sapply(batsBocas_cleandf$cave_group, function(x) {strsplit(x, "-")[[1]][1]})

# remove speed higher than 15 m/s that look like outliers
batsBocas_cleandf <- batsBocas_cleandf[batsBocas_cleandf$ground_speed<=15,]

# rename dataframe
Bocas2023_clean <- batsBocas_cleandf
```

## Down sample data to every 2 minutes
```{r, results=FALSE}
# split by ID
bats_2023_split <- split(Bocas2023_clean, Bocas2023_clean$ID)

# convert to track object before down sampling
resample <-  lapply(bats_2023_split, 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()
})

# convert to data frame
batsdf2023 <- do.call(rbind.data.frame,resample_2)

# I have to do this again, otherwise is a tibble
batsdf2023 <- as.data.frame(batsdf2023)
names(batsdf2023)[1] <- "location_long"
names(batsdf2023)[2] <- "location_lat"
names(batsdf2023)[3] <- "timestamp"

# remove burst columns
batsdf2023 <- batsdf2023[,-54]

batsdf2023[order(batsdf2023$tag_local_identifier, batsdf2023$timestamps),]

# convert to move object
batsmove <- move(x=batsdf2023$location_long, y=batsdf2023$location_lat, 
                 time=as.POSIXct(batsdf2023$timestamps, format="%Y-%m-%d %H:%M:%S", tz="UTC"), 
                 data=batsdf2023, proj=CRS("+proj=longlat +ellps=WGS84"), 
                 animal=batsdf2023$tag_local_identifier, sensor="GPS")

# calculate initial movement parameters pre-cleaning
moveList <- lapply(split(batsmove), 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)
})
n.indiv(batsmove)

movebats2023<- moveStack(moveList, forceTz="UTC")
movebats2023<- spTransform(movebats2023, CRS("+proj=utm +zone=17 +datum=WGS84"))
crds <- as.data.frame(movebats2023@coords)
movebats2023$x <- crds$coords.x1 
movebats2023$y <- crds$coords.x2 

# convert to data frame
batsdf2023_clean <- as.data.frame(movebats2023)

# add id_batday column 
batsdf2023_clean$ID_batday <- paste(batsdf2023_clean$tag_local_identifier, batsdf2023_clean$BatDay, sep="_")
unique(batsdf2023_clean$ID_batday)

# add cave id column
batsdf2023_clean$cave <- sapply(strsplit(batsdf2023_clean$group_id,split = '-'), "[", 1)
# rename la gruta and ajcave from 2023 data
batsdf2023_clean$cave[which(batsdf2023_clean$cave=="LG")] <- "lagruta"
batsdf2023_clean$cave[which(batsdf2023_clean$cave=="MT")] <- "ajcave"

# save data frame cleaned
save(batsdf2023_clean,file="~/ownCloud/PhDLife/P.hastatus/Thesis/Paper1/BiologyLetters/data/PhasBocas_2023_clean.RData")
```