## PROCESSING DATA

# loading packages
pcks <- list('lubridate', 'ggmap', 'maptools', 'rgdal', 'maps', 'mapdata', 'move', 'dplyr', 'geosphere')
sapply(pcks, library, character = TRUE)

# Need to put in username and password below, in quotes
login <- movebankLogin(username=xxxx, password=xxxx)
caribou <- getMovebankData(study="ABoVE: ADFG Fortymile Caribou (for TWS workshop)", login=login)

# If for some reason the line above doesn't work, run this line below:
# load("./_data/caribou.raw.rda")

# converting to a data frame
caribou.df <- as(caribou, "data.frame")
str(caribou.df)

# renaming columns to make them easier to understand
caribou2 <- with(caribou.df, data.frame(id = trackId, lon = location_long, lat = location_lat, time = timestamp, elevation = height_above_ellipsoid))
head(caribou2)

# loading in the buffalo data (to demonstrate working with timestamps from .csv file)
buffalo <- read.csv("./_data/African_buffalo.csv", sep=",", header=TRUE)
str(buffalo)

# converting timestamp to POSIXlt format using two different methods
head(as.POSIXlt(paste(buffalo$timestamp), format = "%Y-%m-%d %H:%M:%S", tz = "UTC"))
head(ymd_hms(buffalo$timestamp))

# the power of the 'lubridate' package!
dates <- c("5/7/1976", "10-24:1979", "1.1.2000")
mdy(dates)
hour(caribou2$time)[1:10]
month(caribou2$time)[1:10]
yday(caribou2$time)[1:10]

# quick plots of the caribou data
require(ggplot2)
ggplot(caribou2, aes(x = lon, y = lat, col = id)) + geom_path()

# determining the correct UTM projection of the caribou data
require(PBSmapping)
ll <- with(caribou2, data.frame(X = lon, Y = lat))
attributes(ll)$projection <- "LL"
xy.km <- convUL(ll, km=TRUE)
str(xy.km)

# calculating x and y ranges of data
diff(range(xy.km$X))
diff(range(xy.km$Y))

# add these transformed coordinates to the caribou data frame
caribou3 <- data.frame(caribou2, x=xy.km$X*1000, y=xy.km$Y*1000)


# what projection is the 'caribou' object?
caribou@proj4string

# projecting to utm
projection <- CRS("+proj=utm +zone=6 +ellps=WGS84 +datum=WGS84 +units=m +no_defs")
caribou.utm <- spTransform(caribou, projection)
plot(caribou.utm)

# converting to a data frame
caribou.utm.df <- as(caribou.utm, "data.frame")
names(caribou.utm.df)

# adding new columns to data frame for x and y (projected coordinates)
xy.projected <- caribou.utm.df[,c("location_long.1", "location_lat.1")]

# seeing if the two projections agree
diff(range(xy.km[,1]))
diff(range(xy.projected[,1]))/1000
diff(range(xy.km[,2]))
diff(range(xy.projected[,2]))/1000

# using dplyr's piping to create a summary file
require(dplyr)
caribou3 %>%
  group_by(id) %>% 
  summarize(
    n.obs = length(id), 
		start = as.Date(min(time)),
		end = as.Date(max(time)),
		duration = max(time) - min(time),
		median.Dt = median(diff(time)),
		range.x = diff(range(x)),
		range.y = diff(range(y))
  )

# two diffeerent plots of temporal extent of data; one for caribou (not interesting), and one for buffalo
ggplot(caribou3, aes(x = time, y = id)) + geom_path()
ggplot(buffalo, aes(x = ymd_hms(timestamp), y = individual.local.identifier)) + geom_path()

# add a column for day.date
caribou3$day.date <- as.Date(caribou3$time)

# create a leaner summarized file
caribou4 <- caribou3 %>% 
  group_by(id, day.date) %>% 
  summarize(
            time = mean(time),
            lon = mean(lon), 
            lat = mean(lat), 
            x = mean(x), y = mean(y),
            elevation = mean(elevation)
  )

# check out the difference between the old file and summarized file
dim(caribou3)
dim(caribou4)


# "one-step" processing

# creating the getUTM function
getUTM <- function(data){
	ll <- with(data, data.frame(X = lon, Y = lat))
	attributes(ll)$projection <- "LL"
	xy <- convUL(ll, km=TRUE)
	data.frame(data, x = xy$X, y = xy$Y)
}

# and doing all the steps above in "one" step!
caribou.daily.df <- caribou %>% 
  as("data.frame") %>%
  dplyr::select(id = trackId, 
         lon = location_long, 
				 lat = location_lat, 
				 time = timestamp,
				 elevation = height_above_ellipsoid) %>%
  mutate(day.date = as.Date(time)) %>% 
  getUTM %>%
  group_by(id, day.date) %>% 
  summarize(
    time = mean(time),
    lon = mean(lon), 
    lat = mean(lat),
    x = mean(x), 
    y = mean(y),
    elevation = mean(elevation)) %>% 
  as.data.frame()

head(caribou.daily.df)