Analysis

Data Analysis

Load Data

library(tidyverse)

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.1.4     ✔ readr     2.1.5
✔ forcats   1.0.0     ✔ stringr   1.5.1
✔ ggplot2   3.5.2     ✔ tibble    3.2.1
✔ lubridate 1.9.4     ✔ tidyr     1.3.1
✔ purrr     1.0.4     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

df_cleaned <- read_csv("cleaned_data.csv")

New names:
Rows: 15786 Columns: 16
── Column specification
──────────────────────────────────────────────────────── Delimiter: "," chr
(10): incident_type, reported_month, region_of_origin, region_of_inciden... dbl
(6): ...1, incident_year, total_death_missing, number_of_females, numbe...
ℹ Use `spec()` to retrieve the full column specification for this data. ℹ
Specify the column types or set `show_col_types = FALSE` to quiet this message.
• `` -> `...1`

Q1 Which migration routes have become the deadliest?

df_cleaned|>
  mutate(migration_route= replace_na(migration_route, "Unknown"))|>
  group_by(migration_route)|>
  summarise(total_deaths = sum(total_death_missing))|>
  arrange(-total_deaths)

# A tibble: 26 × 2
   migration_route                                       total_deaths
   <chr>                                                        <dbl>
 1 Central mediterranean                                        22862
 2 Unknown                                                      10749
 3 Sahara desert crossing                                        8251
 4 Us-mexico border crossing                                     5970
 5 Western africa / atlantic route to the canary islands         4957
 6 Western mediterranean                                         3454
 7 Eastern mediterranean                                         2380
 8 Horn of africa to yemen crossing                              1842
 9 Afghanistan to iran                                           1281
10 Caribbean to us                                                507
# ℹ 16 more rows

Q2 How have migrant deaths changed over the years?

# Step 1: Define Region Groups
df_region <- df_cleaned |>
  mutate(
    # Create a broad Region Group to cluster similar regions together
    `Region Group` = case_when(
      region_of_incident %in% c("North America", "Central America", "South America", "Caribbean") ~ "Americas",
      region_of_incident %in% c("Eastern Asia", "Central Asia", "Western Asia", "Southern Asia", "South-eastern Asia") ~ "Asia",
      region_of_incident %in% c("Northern Africa", "Southern Africa", "Western Africa", "Eastern Africa", "Middle Africa") ~ "Africa",
      region_of_incident %in% c("Europe", "Mediterranean") ~ "Europe & Mediterranean",
      TRUE ~ "Other"),
    # Create a detailed label combining Region Group and specific Region
    `Region Label` = paste(`Region Group`, "-", region_of_incident))
  
# Step 2: Define a Clear Color Palette for Region Groups
region_colors <- c(
  "Africa - Northern Africa" = "#a1d99b",
  "Africa - Southern Africa" = "#74c476",
  "Africa - Western Africa" = "#41ab5d",
  "Africa - Eastern Africa" = "#238b45",
  "Africa - Middle Africa" = "#005a32",
  
  "Americas - North America" = "#fdae6b",
  "Americas - Central America" = "#fd8d3c",
  "Americas - South America" = "#e6550d",
  "Americas - Caribbean" = "#a63603",
  
  "Asia - Eastern Asia" = "#cbc9e2",
  "Asia - Central Asia" = "#9e9ac8",
  "Asia - Western Asia" = "#756bb1",
  "Asia - Southern Asia" = "#54278f",
  "Asia - South-eastern Asia" = "#3f007d",
  
  "Europe & Mediterranean - Europe" = "#fbb4b9",
  "Europe & Mediterranean - Mediterranean" = "#f768a1",
  
  "Other - Other" = "grey60"
)


# Step 3: Group and Plot with facet_wrap()
library(plotly)

Attaching package: 'plotly'

The following object is masked from 'package:ggplot2':

    last_plot

The following object is masked from 'package:stats':

    filter

The following object is masked from 'package:graphics':

    layout

p1 <- df_region|>
  group_by(incident_year, `Region Group`, `Region Label`) |>
  summarise(total_deaths = sum(total_death_missing, na.rm = TRUE), .groups = "drop") |>
  ggplot(aes(
    x = incident_year,
    y = total_deaths,
    group = `Region Label`,
    color = `Region Label`,
    text = paste0("Year: ", incident_year, "<br>Region: ", `Region Label`, "<br>Deaths: ", total_deaths)
  )) +
  geom_line(size = 0.7, alpha = 0.8) +
  geom_point(size = 1.2, alpha = 0.9) +
  facet_wrap(~ `Region Group`, ncol = 2) +
  scale_color_manual(values = region_colors) +
  scale_x_continuous(breaks = seq(2014, max(df_cleaned$incident_year, na.rm = TRUE), 1)) +
  labs(
    title = "Migrant Deaths by Region",
    x = "Year", y = "Total Deaths",
    caption = "Source: IOM Missing Migrants Project | By: Ethan Lim"
  ) +
  theme_minimal(base_size = 9) +
  theme(
    plot.title = element_text(size = 12, face = "bold", hjust = 0.5),
    strip.text = element_text(size = 8, face = "bold"),
    panel.grid.minor = element_blank()
    )

Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.

# Step 4: Make plot interactive
p1 <- ggplotly(p1, tooltip = "text")

# Step 5: Save interactive plot as HTML
install.packages("htmlwidgets")

The following package(s) will be installed:
- htmlwidgets [1.6.4]
These packages will be installed into "~/Desktop/（-S-）/HKU/Year1/Year1_Sem2/JMSC/JMSC1003/Final Project/ethan-website/renv/library/macos/R-4.4/aarch64-apple-darwin20".

# Installing packages --------------------------------------------------------
- Installing htmlwidgets ...                    OK [linked from cache]
Successfully installed 1 package in 4.4 milliseconds.

library(htmlwidgets)

htmlwidgets::saveWidget(p1, "vs/interactive_plot1.html", selfcontained = TRUE)

p1

Q3 What are the most common causes of migrant deaths?

# Tabulate counts for each cause of death
df_cleaned|>
  group_by(cause_of_death)|>
  summarise(n = n())|>
  arrange(-n)

# A tibble: 7 × 2
  cause_of_death             n
  <chr>                  <int>
1 Lack of food and water  4116
2 Drowning                3329
3 Unknown or Mixed        3181
4 Vehicle accident        2114
5 Violence                1316
6 Sickness                1223
7 Accidental death         507

# Step 1: Prepare data for plotting
p2_data <- df_cleaned |>
  filter(!is.na(cause_of_death)) |>
  group_by(cause_of_death) |>
  summarise(deaths = sum(total_death_missing, na.rm = TRUE)) |>
  arrange(-deaths) |>
  slice_max(deaths, n = 10) |>
  mutate(cause_of_death = fct_reorder(cause_of_death, deaths), tooltip = paste("<br>Deaths:", deaths)) 

# Step 2: Create the bar plot
p2 <- ggplot(p2_data, aes(x = cause_of_death, y = deaths, fill = cause_of_death, text = tooltip)) +
  geom_col(show.legend = FALSE, width = 0.7) +
  coord_flip() +
  scale_fill_manual(values = RColorBrewer::brewer.pal(9, "Blues")[3:9]) +
  labs(
    title = "Leading Causes of Migrant Deaths",
    subtitle = "Based on total recorded migrant deaths",
    x = NULL,
    y = "Number of Deaths",
    caption = "Source: IOM Missing Migrants Project | By: Ethan Lim"
  ) +
  theme_minimal(base_family = "Helvetica", base_size = 13) +
  theme(
    plot.title = element_text(size = 16, face = "bold", hjust = 0.5),
    plot.subtitle = element_text(size = 12, hjust = 0.5, margin = margin(b = 10)),
    axis.text.y = element_text(face = "bold"),
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank()
  )

# Step 4: Make plot interactive
p2 <- ggplotly(p2, tooltip = "text")

# Step 5: Save interactive plot as HTML
htmlwidgets::saveWidget(p2, "vs/interactive_plot2.html", selfcontained = TRUE)

p2

Q4 Do migrant deaths follow seasonal patterns?

# Step 1: Create season variable based on reported month
df_cleaned|>
  mutate(season = case_when(
    reported_month == c("December", "January", "February") ~ "Winter",
    reported_month == c("March", "April", "May") ~ "Spring", 
    reported_month == c("June", "July", "August") ~ "Summer", 
    reported_month == c("September", "October", "November") ~ "Autumn",
    TRUE ~ NA
    ))|> 
  filter(!is.na(season))|>
  group_by(season)|>
  summarise(total_deaths = sum(total_death_missing))|>
  arrange(-total_deaths)

# A tibble: 4 × 2
  season total_deaths
  <chr>         <dbl>
1 Summer         6102
2 Autumn         4771
3 Spring         4511
4 Winter         4305

# Step 2: Calculate necessary boxplot statistics manually
stats_df <- df_cleaned |>
  filter(!is.na(reported_month), !is.na(incident_year)) |>
  mutate(
    month = factor(reported_month, levels = month.name, ordered = TRUE),
    year = incident_year
  ) |>
  group_by(month, year) |>
  summarise(deaths = sum(total_death_missing, na.rm = TRUE), .groups = "drop") |>
  group_by(month) |>
  summarise(
    Q1 = quantile(deaths, 0.25, na.rm = TRUE),
    Median = quantile(deaths, 0.5, na.rm = TRUE),
    Q3 = quantile(deaths, 0.75, na.rm = TRUE),
    Min = min(deaths, na.rm = TRUE),
    Max = max(deaths, na.rm = TRUE),
    .groups = "drop"
  )

# Step 3: Create the boxplot
p3 <- df_cleaned |>
  filter(!is.na(reported_month), !is.na(incident_year)) |>
  mutate(
    month = factor(reported_month, levels = month.name, ordered = TRUE),
    year = incident_year
  ) |>
  group_by(month, year) |>
  summarise(deaths = sum(total_death_missing, na.rm = TRUE), .groups = "drop") |>
  ggplot(aes(x = month, y = deaths)) +
  geom_boxplot(
    fill = "#D6EAF8",
    color = "#2980B9",
    outlier.color = "#C0392B",
    outlier.size = 2
  ) +
  # Step 3: Add text labels for Q1, Median, Q3
  geom_text(
    data = stats_df,
    aes(x = month, y = Median, label = paste0("Q2: ", round(Median))),
    vjust = -0.7, size = 2.5, color = "black", inherit.aes = FALSE
  ) +
  geom_text(
    data = stats_df,
    aes(x = month, y = Q1, label = paste0("Q1: ", round(Q1))),
    vjust = 1.5, size = 2.2, color = "darkblue", inherit.aes = FALSE
  ) +
  geom_text(
    data = stats_df,
    aes(x = month, y = Q3, label = paste0("Q3: ", round(Q3))),
    vjust = -1.5, size = 2.2, color = "darkblue", inherit.aes = FALSE
  ) +
  geom_text(
    data = stats_df,
    aes(x = month, y = Min, label = paste0("Min: ", round(Min))),
    vjust = 1.5, size = 2.0, color = "gray40", inherit.aes = FALSE
  ) +
  geom_text(
    data = stats_df,
    aes(x = month, y = Max, label = paste0("Max: ", round(Max))),
    vjust = -1.5, size = 2.0, color = "gray40", inherit.aes = FALSE
  ) +
  labs(
    title = "Distribution of Monthly Migrant Deaths",
    subtitle = "Aggregated by month",
    x = "Month",
    y = "Number of Deaths"
  ) +
  theme_minimal(base_family = "Helvetica") +
  theme(
    plot.title = element_text(size = 13, face = "bold", hjust = 0.5),
    plot.subtitle = element_text(size = 9, hjust = 0.5, margin = margin(b = 10)),
    axis.title = element_text(size = 9),
    axis.text.x = element_text(size = 7, angle = 45, hjust = 1),
    axis.text.y = element_text(size = 7),
    panel.grid.major.y = element_line(color = "gray90"),
    panel.grid.minor = element_blank(),
    panel.grid.major.x = element_blank(),
    plot.background = element_rect(fill = "white", color = NA),
    panel.background = element_rect(fill = "white", color = NA)
  )

# Step 3: Save plot as a PNG file
ggsave(filename = "vs/Q4.png", plot = p3, width = 8, height = 5, dpi = 300)

p3

Q5 Where are migrant deaths geographically concentrated?

# Step 1: Install & Load packages
install.packages("rnaturalearth") # For country shapefiles

The following package(s) will be installed:
- rnaturalearth [1.0.1]
These packages will be installed into "~/Desktop/（-S-）/HKU/Year1/Year1_Sem2/JMSC/JMSC1003/Final Project/ethan-website/renv/library/macos/R-4.4/aarch64-apple-darwin20".

# Installing packages --------------------------------------------------------
- Installing rnaturalearth ...                  OK [linked from cache]
Successfully installed 1 package in 3.1 milliseconds.

install.packages("rnaturalearthdata") # Dependency for rnaturalearth

The following package(s) will be installed:
- rnaturalearthdata [1.0.0]
These packages will be installed into "~/Desktop/（-S-）/HKU/Year1/Year1_Sem2/JMSC/JMSC1003/Final Project/ethan-website/renv/library/macos/R-4.4/aarch64-apple-darwin20".

# Installing packages --------------------------------------------------------
- Installing rnaturalearthdata ...              OK [linked from cache]
Successfully installed 1 package in 2.9 milliseconds.

library(rnaturalearth)

# Step 2: Extract latitude and longitude from 'coordinates'
df_cleaned <- df_cleaned |>
  separate(coordinates, into = c("latitude", "longitude"), sep = ",", remove = FALSE) |>
  mutate(
    latitude = as.numeric(trimws(latitude)),
    longitude = as.numeric(trimws(longitude))
  )

# Step 3: Filter valid points (non-missing coordinates and deaths > 0)
df_points <- df_cleaned |>
  filter(!is.na(latitude), !is.na(longitude), total_death_missing > 0)

# Step 4: Load world map
world <- ne_countries(scale = "medium", returnclass = "sf")

# Step 5: Plot map
map0 <- ggplot() +
  geom_sf(data = world, fill = "gray95", color = "gray70", size = 0.2) +
  geom_point(
    data = df_points,
    aes(x = longitude, y = latitude),
    color = "red",
    fill = "white",
    size = 1.2,
    alpha = 0.7
  ) +
  labs(
    title = "Geographic Locations of Migrant Deaths",
    subtitle = "Each red dot marks a recorded death location",
    caption = "Source: Missing Migrants Project (IOM) | By: Ethan Lim"
  ) +
  coord_sf(xlim = c(-180, 180), ylim = c(-60, 80), expand = FALSE) +
  theme_minimal() +
  theme(
    plot.title = element_text(size = 13, face = "bold"),
    plot.subtitle = element_text(size = 10),
    plot.caption = element_text(size = 8),
    legend.position = "none",
    panel.background = element_rect(fill = "white", color = NA),
    plot.background = element_rect(fill = "white", color = NA)
  )

# Step 5: Save the map as PNG file
ggsave(filename = "vs/Q6.png", plot = map0, width = 10, height = 6, dpi = 300)

map0

Q6 How do migrant deaths vary by region of origin?

# Step 1: Install & Load packages
install.packages("RColorBrewer") # For color palettes

The following package(s) will be installed:
- RColorBrewer [1.1-3]
These packages will be installed into "~/Desktop/（-S-）/HKU/Year1/Year1_Sem2/JMSC/JMSC1003/Final Project/ethan-website/renv/library/macos/R-4.4/aarch64-apple-darwin20".

# Installing packages --------------------------------------------------------
- Installing RColorBrewer ...                   OK [linked from cache]
Successfully installed 1 package in 3.3 milliseconds.

library(RColorBrewer)


# Step 2: Summarize deaths by country of origin
df_country <- df_cleaned|>
  filter(country_of_origin != "Unknown", !is.na(country_of_origin))|>
  group_by(country_of_origin)|>
  summarise(total = sum(total_death_missing))

# Step 3: Load World Map
world_map <- ne_countries(scale = "medium", returnclass = "sf")|>
  select(name_long, geometry)

# Step 4: Merge deaths to map by country name
world_map_deaths <- world_map|>
  left_join(df_country, by = c("name_long" = "country_of_origin"))

# Step 5: Plot map

map1 <- ggplot(world_map_deaths) +
  geom_sf(aes(fill = total), color = "gray50", size = 0.1) +
  scale_fill_distiller(
    palette = "Blues",
    na.value = "white", 
    trans = "log",    
    direction = 1,
    name = "Total Deaths",
    breaks = c(1,10,100,1000,2000),
    guide = guide_colorbar(barwidth = 18, barheight = 0.8, title.position = "top", title.hjust = 0.5)) +
  labs(
    title = "Distribution of Migrant Deaths",
    subtitle = "By Region of Origin",
    caption = "Source: Missing Migrants Project (IOM) | By: Ethan Lim"
  ) +
  theme_minimal() +
  theme(
    plot.title = element_text(size = 12, face = "bold"),
    plot.subtitle = element_text(size = 9),
    legend.position = "bottom",
    legend.title = element_text(size = 10),
    legend.text = element_text(size = 8))
  
# Step 6: Save the map as a PNG file
ggsave(filename = "vs/Q5.png", plot = map1, width = 10, height = 6, dpi = 300)

map1