---
title: "Summarising Tracking Data"
author: "Matt Crane, Inês Silva, Benjamin Michael Marshall, Colin Thomas Strine"
date: "24/08/2020"
output:
  pdf_document: default
  html_document: default
---

```{r echo = FALSE, message=FALSE}

library(dplyr)
library(ggplot2)

animalData <- read.csv(file = "./Supp File 3 - King cobra data.csv", 
                       stringsAsFactors = FALSE)

animalData <- animalData %>% 
  group_by(individual.local.identifier) %>% 
  mutate(study.local.timestamp = as.POSIXct(study.local.timestamp,
                                            tz = "Asia/Bangkok")) %>% 
  arrange(study.local.timestamp) %>% 
  mutate(step.length = sqrt((utm.easting - lag(utm.easting))^2 +
                              (utm.northing - lag(utm.northing))^2),
         time.lag = as.numeric(difftime(study.local.timestamp, lag(study.local.timestamp),
                             units = "hours")))

summaryTable <- animalData %>% 
  group_by(individual.local.identifier) %>% 
  summarise(datapoints = length(timestamp),
            duration = as.numeric(difftime(max(study.local.timestamp),
                                           min(study.local.timestamp),
                                             units = "days")),
            timelag = paste0(round(digits = 2, mean(time.lag, na.rm = TRUE)), " ±",
                             round(digits = 2,
                                   sd(time.lag, na.rm = TRUE)/
                                     sqrt(length(time.lag)))),
            moves = sum(step.length > 0, na.rm = TRUE),
            step.length = paste0(round(digits = 2, mean(step.length, na.rm = TRUE)), 
                                 " ±", round(digits = 2, sd(step.length, na.rm = TRUE)/
                                               sqrt(length(step.length))))
            )

intextSummaries <- summaryTable %>% 
  summarise(datapointsMean = mean(datapoints),
            datapointsSE = sd(datapoints)/sqrt(length(datapoints)),
            movesMean = mean(moves),
            movesSE = sd(moves)/sqrt(length(moves)),
            durationMean = mean(duration),
            durationSE = sd(duration)/sqrt(length(duration))
            )

overallSummaries <- animalData %>% 
  ungroup() %>% 
  summarise(timelagMean = mean(time.lag, na.rm = TRUE),
            timelagSE = sd(time.lag, na.rm = TRUE)/sqrt(length(time.lag)),
            timelagMax = max(time.lag, na.rm = TRUE),
            timelagMin = min(time.lag, na.rm = TRUE),
            steplengthMean = mean(step.length, na.rm = TRUE),
            steplengthSE = sd(step.length, na.rm = TRUE)/sqrt(length(step.length))
            )

```

We tracked `r length(unique(animalData$individual.local.identifier))` [SPECIES NAME] individuals for an average of `r round(intextSummaries$durationMean, digits = 2)` ±`r round(intextSummaries$durationSE, digits = 2)` days. During the tracking period, we located individuals on average every `r round(overallSummaries$timelagMean, digits = 2)` ±`r round(overallSummaries$timelagSE, digits = 2)` hours (mean time lag), and detected `r round(intextSummaries$movesMean, digits = 2)` ±`r round(intextSummaries$movesSE, digits = 2)` moves per individual with a mean step length of `r round(overallSummaries$steplengthMean, digits = 2)` ±`r round(overallSummaries$steplengthSE, digits = 2)` m.

```{r echo=FALSE}
knitr::kable(summaryTable, digits = 2,
             caption = "Summary of animal tracking",
             col.names = c("Animal ID", "# datapoints", 
                           "Duration (days)",
                           "Mean time lag ±SE (hours)",
                           "# moves",
                           "Mean step length ±SE (m)"
                           ),
             align = "c")
```

```{r echo=FALSE, fig.align='center', fig.cap="The tracking duration of all indivudals.", fig.dim=c(12, 4)}

ggplot(animalData) +
  geom_point(aes(x = study.local.timestamp, y = individual.local.identifier,
                 colour = time.lag > 12 & !is.na(time.lag)),
             size = 1, shape = "|") +
  scale_colour_manual(values = c("grey10", "darkred"),
                      labels = c("< 12 hour time lag",
                                 "> 12 hour time lag")) +
  labs(x = "Date", y = "Animal ID", colour = "Time lag larger\nthan planned") +
  theme_bw() +
  theme(axis.title = element_text(face = 2),
        axis.title.y = element_text(hjust = 0, angle = 0)) +
  guides(colour = guide_legend(override.aes = list(size = 5)))


```

```{r echo=FALSE, message=FALSE, warning=FALSE, fig.align='center', fig.cap="The distribution and mean (dashed line) lag time between tracks. Note, x-axis is square-rooted and truncated at 96 hours", fig.dim=c(8, 4)}

animalData %>% 
  filter(!is.na(time.lag)) %>% 
  ggplot() +
  geom_vline(aes(xintercept = mean(time.lag)),
             linetype = 2, alpha = 0.5) +
  geom_density(aes(x = time.lag), fill = "black", alpha = 0.5, colour = NA) +
  annotate("text", x = 96, y = 0.85,
           label = paste0("Mean lag = ", 
                          round(digits = 2, overallSummaries$timelagMean),
                          " ± ", 
                          round(digits = 2, overallSummaries$timelagSE),
                          " hours\n",
                          "Range = ", 
                          round(digits = 2, overallSummaries$timelagMin),
                          " - ", 
                          round(digits = 2, overallSummaries$timelagMax), " hours"),
           fontface = 4, hjust = 1) +
  theme_bw() +
  labs(x = "Time lag between tracks (hours)", y = "Density") +
  scale_x_sqrt(breaks = c(2, seq(0,12,6), seq(24, 120, 24),
                          seq(240, 800, 240)),
               limits = c(0, 96)) +
  theme(axis.title.y = element_text(angle = 0, vjust = 1, face = 2),
        axis.title.x = element_text(hjust = 1, face = 2))

```