caribou-locations.Rmd

---
title: "Untitled"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r}
library(tidyverse)
theme_set(theme_light())

tuesdata <- tidytuesdayR::tt_load('2020-06-23')
```

```{r}
individuals <- tuesdata$individuals
locations <- tuesdata$locations
```

```{r}
individuals %>%
  summarize(across(sex:study_site, list(~ mean(!is.na(.)))))

individuals %>%
  filter(deploy_off_type == "dead") %>%
  count(death_cause, sort = TRUE)
```

```{r}
library(sf)
province_sf <- read_sf("~/Downloads/province")

bc <- province_sf %>%
  filter(PROV == "BC")
```

```{r}
individuals %>%
  filter(deploy_on_latitude > 40) %>%
  count(study_site, deploy_on_longitude, deploy_on_latitude, sort = TRUE) %>%
  ggplot() +
  # geom_sf(data = bc) +
  # borders("world", regions = "canada") +
  geom_point(aes(deploy_on_longitude, deploy_on_latitude,
                 size = n,
             color = study_site)) +
  scale_size_continuous(guide = FALSE)
```

```{r}
individuals %>%
  count(animal_id, sort = TRUE)

individuals %>%
  filter(animal_id == "MO_car150") %>%
  View()
```

### Looking at locations

```{r}
locations %>%
  ggplot(aes(longitude, latitude, color = study_site)) +
  geom_point()
```

```{r}
by_animal <- locations %>%
  group_by(animal_id, study_site) %>%
  summarize(start = min(timestamp),
            end = max(timestamp),
            num_points = n()) %>%
  ungroup() %>%
  arrange(desc(num_points))
```

```{r}
library(lubridate)

example_animal <- locations %>%
  arrange(timestamp) %>%
  filter(animal_id == sample(unique(animal_id), 1))

example_animal %>%
  mutate(quarter = as.Date(floor_date(timestamp, "quarter"))) %>%
  ggplot(aes(longitude, latitude, color = timestamp)) +
  geom_point(alpha = .5) +
  geom_path(alpha = .5) +
  facet_wrap(~ quarter) +
  labs(title = "One caribou over time")
```

```{r}
library(geosphere)

locations_with_gaps <- locations %>%
  group_by(animal_id) %>%
  mutate(last_lon = lag(longitude),
         last_lat = lag(latitude),
         hours = as.numeric(difftime(timestamp, lag(timestamp), unit = "hours")),
         km = distHaversine(cbind(longitude, latitude), cbind(last_lon, last_lat)) / 1000,
         kph = km / hours) %>%
  ungroup()

locations_with_gaps %>%
  filter(hours <= 8) %>%
  ggplot(aes(kph)) +
  geom_histogram() +
  scale_x_log10(labels = scales::comma) +
  labs(title = "On average, how fast do caribou travel?")

by_animal <- locations_with_gaps %>%
  filter(hours <= 8,
         hours >= .5) %>%
  group_by(animal_id, study_site) %>%
  summarize(start = min(timestamp),
            end = max(timestamp),
            num_points = n(),
            avg_speed = mean(kph, na.rm = TRUE)) %>%
  ungroup() %>%
  arrange(desc(num_points)) %>%
  filter(num_points >= 10)
```

```{r}
by_animal %>%
  ggplot(aes(num_points, avg_speed)) +
  geom_point() +
  scale_x_log10() +
  expand_limits(y = 0)

by_animal %>%
  arrange(desc(avg_speed))

locations_with_gaps %>%
  filter(animal_id == "QU_car107") %>%
  mutate(quarter = as.Date(floor_date(timestamp, "quarter"))) %>%
  ggplot(aes(longitude, latitude, color = kph)) +
  geom_point(alpha = .5) +
  geom_path(alpha = .5) +
  facet_wrap(~ quarter) +
  labs(title = "One caribou over time")

locations_with_gaps %>%
  filter(animal_id == "SC_car171") %>%
  arrange(desc(kph)) %>%
  View()
```

```{r}
locations_with_gaps %>%
  filter(study_site != "Hart Ranges") %>%
  filter(hours <= 8,
         hours >= .5) %>%
  group_by(month = month(timestamp, label = TRUE),
           study_site) %>%
  summarize(avg_speed = median(kph),
            n = n()) %>%
  ggplot(aes(month, avg_speed, group = study_site,
             color = study_site)) +
  geom_line() +
  geom_point(aes(size = n)) +
  expand_limits(y = 0) +
  facet_wrap(~ study_site) +
  theme(legend.position = "none") +
  labs(title = "Seasonal trend in Caribou speed",
       y = "Average speed (kph)")
```


```{r}
by_animal %>%
  filter(num_points >= 10) %>%
  arrange(desc(avg_speed))
```


```{r}
locations %>%
  arrange(timestamp) %>%
  group_by(animal_id) %>%
  mutate(gap = round(difftime(timestamp, lag(timestamp), unit = "hours"))) %>%
  ungroup() %>%
  filter(!is.na(gap)) %>%
  filter(gap <= 24) %>%
  ggplot(aes(gap)) +
  geom_histogram(binwidth = 2)
```