Exploring Patterns and Trends in Terrorism Impact
This analysis explores the Global Terrorism Index (GTI) data from 2012 to 2022, examining trends and patterns in terrorism impact across countries. The GTI is a comprehensive measure that incorporates incidents, fatalities, injuries, and hostage situations to quantify the impact of terrorism globally.
The first visualization shows the overall trend of GTI scores for the top 15 most affected countries from 2012 to 2022. This provides a comparative view of how terrorism impact has evolved over time across these nations.
# Plot the data
ggplot(filtered_df, aes(x = Year, y = Score, color = Country)) +
geom_line(linewidth = 1.2) + # ← updated here
geom_point() +
theme_minimal() +
labs(title = "Top 15 Countries by Terrorism Index Score (2012–2022)",
y = "GTI Score",
x = "Year",
color = "Country") +
theme(plot.title = element_text(hjust = 0.5, face = "bold"),
legend.position = "bottom",
legend.title = element_text(face = "bold"))
Key Insights:
• Afghanistan has consistently maintained the highest GTI score throughout most of the period
• Iraq shows a significant decline in terrorism impact since its peak around 2014-2015
• Several countries show recent increases in their GTI scores, suggesting emerging terrorism hotspots
• The overall pattern reveals fluctuating impacts, indicating the dynamic nature of terrorism threats globally
This visualization breaks down the terrorism index trends by country with clear severity classifications. The background color bands indicate different severity levels (Severe, High, Moderate, Low), providing context for interpreting the GTI scores.
ggplot(filtered_df, aes(x = Year, y = Score)) +
# Severity background bands (more visible)
annotate("rect", xmin = -Inf, xmax = Inf, ymin = 8, ymax = 10,
fill = "#B22222", alpha = 0.15) +
annotate("rect", xmin = -Inf, xmax = Inf, ymin = 6, ymax = 8,
fill = "#FF8C00", alpha = 0.15) +
annotate("rect", xmin = -Inf, xmax = Inf, ymin = 4, ymax = 6,
fill = "#FFD700", alpha = 0.15) +
annotate("rect", xmin = -Inf, xmax = Inf, ymin = 0, ymax = 4,
fill = "#90EE90", alpha = 0.15) +
# Add labels to indicate severity zones
annotate("text", x = 2013, y = 9, label = "Severe", color = "#8B0000", size = 4, fontface = "bold") +
annotate("text", x = 2013, y = 7, label = "High", color = "#FF8C00", size = 4, fontface = "bold") +
annotate("text", x = 2013, y = 5, label = "Moderate", color = "#DAA520", size = 4, fontface = "bold") +
annotate("text", x = 2013, y = 2, label = "Low", color = "#228B22", size = 4, fontface = "bold") +
# Main line chart
geom_line(color = "#000000", linewidth = 1.2) +
geom_point(color = "#000000", size = 1.2) +
# Faceted by country
facet_wrap(~Country, ncol = 5) +
theme_minimal(base_size = 12) +
labs(
title = "Terrorism Index Trends by Country (2012–2022)",
subtitle = "GTI Scores Categorized by Severity Levels",
x = "Year",
y = "GTI Score (0–10)\nHigher = More Impact from Terrorism"
) +
theme(
plot.title = element_text(size = 16, face = "bold", hjust = 0.5),
plot.subtitle = element_text(size = 13, hjust = 0.5),
strip.text = element_text(size = 11, face = "bold"),
axis.text.x = element_text(angle = 45, hjust = 1),
panel.grid.minor = element_blank()
)
Key Insights:
• Most countries have experienced periods in the “High” to “Severe” range during the decade
• Some countries show clear improvement patterns (Iraq, Syria, Nigeria)
• Others demonstrate concerning upward trends (Burkina Faso, Mali)
• Afghanistan has remained consistently in the “Severe” category
• The severity classification provides important context for understanding the relative impact of terrorism
This visualization shows the countries with the highest number of terrorism-related fatalities over the 2012-2022 period. The number of fatalities is a direct measure of the human cost of terrorism.
top_deaths <- combined_raw_df %>%
group_by(Country) %>%
summarise(Total_Deaths = sum(Fatalities, na.rm = TRUE)) %>%
arrange(desc(Total_Deaths)) %>%
slice_head(n = 15)
ggplot(top_deaths, aes(x = reorder(Country, Total_Deaths), y = Total_Deaths)) +
geom_col(fill = "darkred") +
# Add value labels at the end of each bar
geom_text(aes(label = format(Total_Deaths, big.mark = ",")),
hjust = -0.2, color = "black", fontface = "bold", size = 3.5) +
coord_flip(clip = "off") + # Prevent clipping of labels
scale_y_continuous(limits = function(x) c(0, x[2] * 1.1)) + # Expand y-axis to make room for labels
labs(title = "Top 15 Countries by Total Fatalities (2012–2022)",
x = "Country", y = "Total Fatalities") +
theme_minimal() +
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
axis.text.y = element_text(face = "bold"),
plot.margin = margin(5.5, 30, 5.5, 5.5) # Add right margin for labels
)
Beyond fatalities, injuries represent another significant impact of terrorism. This visualization shows the countries with the highest number of terrorism-related injuries, which often outnumber fatalities and create long-term burdens on healthcare systems.
top_injuries <- combined_raw_df %>%
group_by(Country) %>%
summarise(Total_Injuries = sum(Injuries, na.rm = TRUE)) %>%
arrange(desc(Total_Injuries)) %>%
slice_head(n = 15)
ggplot(top_injuries, aes(x = reorder(Country, Total_Injuries), y = Total_Injuries)) +
geom_col(fill = "orange") +
# Add value labels at the end of each bar
geom_text(aes(label = format(Total_Injuries, big.mark = ",")),
hjust = -0.2, color = "black", fontface = "bold", size = 3.5) +
coord_flip(clip = "off") + # Prevent clipping of labels
scale_y_continuous(limits = function(x) c(0, x[2] * 1.1)) + # Expand y-axis to make room for labels
labs(title = "Top 15 Countries by Total Injuries (2012–2022)",
x = "Country", y = "Total Injuries") +
theme_minimal() +
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
axis.text.y = element_text(face = "bold"),
plot.margin = margin(5.5, 30, 5.5, 5.5) # Add right margin for labels
)
The number of terrorist incidents provides a measure of the frequency of attacks, regardless of their scale. This visualization highlights countries experiencing the highest number of terrorism events during the study period.
top_incidents <- combined_raw_df %>%
group_by(Country) %>%
summarise(Total_Incidents = sum(Incidents, na.rm = TRUE)) %>%
arrange(desc(Total_Incidents)) %>%
slice_head(n = 15)
ggplot(top_incidents, aes(x = reorder(Country, Total_Incidents), y = Total_Incidents)) +
geom_col(fill = "steelblue") +
# Add value labels at the end of each bar
geom_text(aes(label = format(Total_Incidents, big.mark = ",")),
hjust = -0.2, color = "black", fontface = "bold", size = 3.5) +
coord_flip(clip = "off") + # Prevent clipping of labels
scale_y_continuous(limits = function(x) c(0, x[2] * 1.1)) + # Expand y-axis to make room for labels
labs(title = "Top 15 Countries by Total Incidents (2012–2022)",
x = "Country", y = "Total Incidents") +
theme_minimal() +
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
axis.text.y = element_text(face = "bold"),
plot.margin = margin(5.5, 30, 5.5, 5.5) # Add right margin for labels
)
top_hostages <- combined_raw_df %>%
group_by(Country) %>%
summarise(Total_Hostages = sum(Hostages, na.rm = TRUE)) %>%
arrange(desc(Total_Hostages)) %>%
slice_head(n = 15)
ggplot(top_hostages, aes(x = reorder(Country, Total_Hostages), y = Total_Hostages)) +
geom_col(fill = "purple") +
# Add value labels at the end of each bar
geom_text(aes(label = format(Total_Hostages, big.mark = ",")),
hjust = -0.2, color = "black", fontface = "bold", size = 3.5) +
coord_flip(clip = "off") + # Prevent clipping of labels
scale_y_continuous(limits = function(x) c(0, x[2] * 1.1)) + # Expand y-axis to make room for labels
labs(title = "Top 15 Countries by Total Hostages (2012–2022)",
x = "Country", y = "Total Hostages Taken") +
theme_minimal() +
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
axis.text.y = element_text(face = "bold"),
plot.margin = margin(5.5, 30, 5.5, 5.5) # Add right margin for labels
)
Key Insights from Impact Analysis:
• Fatalities: Iraq, Afghanistan, and Nigeria have suffered the highest number of terrorism-related deaths
• Injuries: Iraq and Afghanistan also lead in injuries, but Pakistan ranks higher in injuries than in fatalities
• Incidents: Iraq experiences the most terrorism incidents, followed by Afghanistan and Pakistan
• Hostages: Nigeria stands out with significantly more hostages than other countries, likely due to specific tactics used by terrorist groups operating there
• Different countries show different patterns of terrorism impact (e.g., some have high fatality-to-incident ratios while others show more incidents with fewer casualties)
This visualization tracks the global yearly trends in terrorism impact, showing how fatalities, injuries, and hostage situations have evolved over the decade. This provides a macro-level view of global terrorism patterns.
global_trend <- combined_raw_df %>%
group_by(Year) %>%
summarise(Fatalities = sum(Fatalities, na.rm = TRUE),
Injuries = sum(Injuries, na.rm = TRUE),
Hostages = sum(Hostages, na.rm = TRUE))
global_trend_long <- pivot_longer(global_trend,
cols = c("Fatalities", "Injuries", "Hostages"),
names_to = "Type", values_to = "Count")
ggplot(global_trend_long, aes(x = Year, y = Count, color = Type)) +
geom_line(size = 1.2) +
labs(title = "Global Trends: Fatalities, Injuries, Hostages (2012–2022)",
x = "Year", y = "Count", color = "Type") +
theme_minimal()Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
ℹ Please use `linewidth` instead.
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
legend.position = "bottom",
legend.title = element_text(face = "bold")
)List of 3
$ legend.title :List of 11
..$ family : NULL
..$ face : chr "bold"
..$ colour : NULL
..$ size : NULL
..$ hjust : NULL
..$ vjust : NULL
..$ angle : NULL
..$ lineheight : NULL
..$ margin : NULL
..$ debug : NULL
..$ inherit.blank: logi FALSE
..- attr(*, "class")= chr [1:2] "element_text" "element"
$ legend.position: chr "bottom"
$ plot.title :List of 11
..$ family : NULL
..$ face : chr "bold"
..$ colour : NULL
..$ size : NULL
..$ hjust : num 0.5
..$ vjust : NULL
..$ angle : NULL
..$ lineheight : NULL
..$ margin : NULL
..$ debug : NULL
..$ inherit.blank: logi FALSE
..- attr(*, "class")= chr [1:2] "element_text" "element"
- attr(*, "class")= chr [1:2] "theme" "gg"
- attr(*, "complete")= logi FALSE
- attr(*, "validate")= logi TRUEKey Global Trends:
• The period around 2014-2015 shows a significant peak in both fatalities and injuries, corresponding to the rise of ISIS
• Since 2015, there has been a general downward trend in all three impact measures
• Hostage situations follow a different pattern than fatalities and injuries, with distinct peaks
• Recent years show some stabilization at lower levels compared to the mid-decade peak
This bubble chart provides a multi-dimensional view of terrorism impact, plotting fatalities against injuries while using bubble size to represent the average GTI score. This visualization helps identify countries that are disproportionately affected across different impact measures.
bubble_data <- combined_raw_df %>%
group_by(Country) %>%
summarise(Fatalities = sum(Fatalities, na.rm = TRUE),
Injuries = sum(Injuries, na.rm = TRUE),
GTI = mean(Score, na.rm = TRUE)) %>%
filter(Fatalities > 0 & Injuries > 0)
ggplot(bubble_data, aes(x = Fatalities, y = Injuries, size = GTI, label = Country)) +
geom_point(alpha = 0.6, color = "darkblue") +
geom_text(check_overlap = TRUE, size = 3, vjust = 1.5) +
scale_size_continuous(range = c(2, 10)) +
labs(title = "Bubble Chart: Fatalities vs Injuries by Country",
x = "Total Fatalities", y = "Total Injuries", size = "Avg. GTI Score") +
theme_minimal()
theme(
plot.title = element_text(face = "bold", hjust = 0.5),
plot.subtitle = element_text(hjust = 0.5),
legend.position = "right"
)List of 3
$ legend.position: chr "right"
$ plot.title :List of 11
..$ family : NULL
..$ face : chr "bold"
..$ colour : NULL
..$ size : NULL
..$ hjust : num 0.5
..$ vjust : NULL
..$ angle : NULL
..$ lineheight : NULL
..$ margin : NULL
..$ debug : NULL
..$ inherit.blank: logi FALSE
..- attr(*, "class")= chr [1:2] "element_text" "element"
$ plot.subtitle :List of 11
..$ family : NULL
..$ face : NULL
..$ colour : NULL
..$ size : NULL
..$ hjust : num 0.5
..$ vjust : NULL
..$ angle : NULL
..$ lineheight : NULL
..$ margin : NULL
..$ debug : NULL
..$ inherit.blank: logi FALSE
..- attr(*, "class")= chr [1:2] "element_text" "element"
- attr(*, "class")= chr [1:2] "theme" "gg"
- attr(*, "complete")= logi FALSE
- attr(*, "validate")= logi TRUEKey Insights from Multi-Dimensional Analysis:
• Iraq and Afghanistan stand as extreme outliers, with exceptionally high values across all three dimensions
• Most countries cluster in the lower ranges for both fatalities and injuries
• Some countries show disproportionate ratios between fatalities and injuries
• The relationship between GTI score and casualty figures is generally consistent, with higher casualties correlating with higher GTI scores
• A few countries have relatively high GTI scores despite moderate casualty figures, suggesting other factors (like property damage or psychological impact) influence their GTI ranking
Key Metric: Fatality Rate
• Definition: Average deaths per attack = Fatalities ÷ Incidents.
• Why it matters: Normalizes for attack frequency so you can compare how deadly incidents are across countries and over time.
• Interpretation: An increasing rate signals more severe attacks or reduced survivability; a declining rate suggests less-lethal tactics or improved emergency response.
Data Ingestion: Load the raw “Combined raw” sheet into R using read_excel, preserving the original structure so we can inspect and verify the incoming data before any transformations.
library(readxl)
data_path <- "data/Global Terrorism Index 2023.xlsx"
raw_df <- read_excel(data_path, sheet = "Combined raw")
head(raw_df)# A tibble: 6 × 9
iso3c Country Rank Score Incidents Fatalities Injuries Hostages Year
<chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 IRQ Iraq 1 9.60 1288 2086 5050 16 2012
2 PAK Pakistan 2 9.15 638 1322 2297 160 2012
3 AFG Afghanistan 3 9.13 507 1511 2612 67 2012
4 SYR Syria 4 8.24 168 1014 1833 71 2012
5 YEM Yemen 5 8.10 219 651 798 121 2012
6 NGA Nigeria 6 8.10 212 767 496 9 2012Initial Data Snapshot:
- The raw “Combined raw” sheet contains 9 columns: iso3c, Country, Rank, Score, Incidents, Fatalities, Injuries, Hostages, and Year.
- The first six rows for 2012 list Iraq, Pakistan, Afghanistan, Syria, Yemen, and Nigeria—showing their 2012 GTI rank, score, and casualty figures.
- Confirms successful ingestion and expected structure before any cleaning or transformation.
Cleaning & Normalization: Standardize column names to snake_case, remove empty or “Total” rows, trim whitespace in country, convert all metric columns (incidents, fatalities, injuries, hostages, score) to numeric types, replace any missing values with zeros, and drop exact duplicate rows to ensure a consistent dataset.
library(dplyr)
library(janitor)
Attaching package: 'janitor'The following objects are masked from 'package:stats':
chisq.test, fisher.testlibrary(stringr)
clean_df <- raw_df %>%
clean_names() %>% # snake_case column names
filter(
!is.na(country),
country != "",
!str_detect(country, regex("total|all", ignore_case = TRUE))
) %>%
mutate(
country = str_squish(country), # trim extra spaces
year = as.integer(year),
incidents = as.numeric(incidents),
fatalities = as.numeric(fatalities),
injuries = as.numeric(injuries),
hostages = as.numeric(hostages),
score = as.numeric(score)
) %>%
replace_na(list( # missing counts → 0
incidents = 0,
fatalities = 0,
injuries = 0,
hostages = 0
)) %>%
distinct(country, year, .keep_all = TRUE) # drop exact duplicates
head(clean_df)# A tibble: 6 × 9
iso3c country rank score incidents fatalities injuries hostages year
<chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <int>
1 IRQ Iraq 1 9.60 1288 2086 5050 16 2012
2 PAK Pakistan 2 9.15 638 1322 2297 160 2012
3 AFG Afghanistan 3 9.13 507 1511 2612 67 2012
4 SYR Syria 4 8.24 168 1014 1833 71 2012
5 YEM Yemen 5 8.10 219 651 798 121 2012
6 NGA Nigeria 6 8.10 212 767 496 9 2012Validate Cleaning Results: Verify that column types, row counts, and missing‐value replacements are as expected. We compare before/after dimensions, check for duplicates, and summarize any remaining NAs.
library(dplyr)
# 1. Compare dimensions before & after
cat("Raw rows:", nrow(raw_df), "→ Clean rows:", nrow(clean_df), "\n")Raw rows: 1793 → Clean rows: 1793 cat("Raw cols:", ncol(raw_df), "→ Clean cols:", ncol(clean_df), "\n\n")Raw cols: 9 → Clean cols: 9 # 2. Check for any duplicate rows remaining
cat("Duplicates in clean_df:", sum(duplicated(clean_df)), "\n\n")Duplicates in clean_df: 0 # 3. Summarize NA counts by column
na_summary <- clean_df %>%
summarise(across(everything(), ~ sum(is.na(.))))
print(na_summary)# A tibble: 1 × 9
iso3c country rank score incidents fatalities injuries hostages year
<int> <int> <int> <int> <int> <int> <int> <int> <int>
1 0 0 0 0 0 0 0 0 0# 4. Glimpse the cleaned structure
glimpse(clean_df)Rows: 1,793
Columns: 9
$ iso3c <chr> "IRQ", "PAK", "AFG", "SYR", "YEM", "NGA", "SOM", "IND", "TH…
$ country <chr> "Iraq", "Pakistan", "Afghanistan", "Syria", "Yemen", "Niger…
$ rank <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, …
$ score <dbl> 9.599967, 9.152620, 9.134265, 8.238079, 8.098513, 8.098342,…
$ incidents <dbl> 1288, 638, 507, 168, 219, 212, 166, 219, 199, 219, 157, 119…
$ fatalities <dbl> 2086, 1322, 1511, 1014, 651, 767, 446, 150, 184, 157, 188, …
$ injuries <dbl> 5050, 2297, 2612, 1833, 798, 496, 427, 256, 493, 309, 502, …
$ hostages <dbl> 16, 160, 67, 71, 121, 9, 1, 41, 0, 3, 89, 21, 29, 1, 6, 3, …
$ year <int> 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012, 2012,…Post-Cleaning Validation:
• Row & Column Consistency: Retained all 1,793 rows and 9 columns—no unintended data loss.
• Duplicates Removed: Zero duplicate entries, ensuring each country–year is unique.
• No Missing Values: All incidents, fatalities, injuries, and hostages fields now have zero NAs.
• Correct Data Types: year is integer and all metric columns are numeric, ready for reliable aggregation and analysis.
Filtering & Period Selection: Restrict the cleaned dataset to the years 2012–2022 (inclusive), so that all subsequent analysis focuses on our target time window.
# 1. Keep only 2012 through 2022
combined_raw_df <- clean_df %>%
filter(year >= 2012, year <= 2022)
# 2. Verify the operation
cat("Year range after filter:", range(combined_raw_df$year), "\n")Year range after filter: 2012 2022 head(combined_raw_df)# A tibble: 6 × 9
iso3c country rank score incidents fatalities injuries hostages year
<chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <int>
1 IRQ Iraq 1 9.60 1288 2086 5050 16 2012
2 PAK Pakistan 2 9.15 638 1322 2297 160 2012
3 AFG Afghanistan 3 9.13 507 1511 2612 67 2012
4 SYR Syria 4 8.24 168 1014 1833 71 2012
5 YEM Yemen 5 8.10 219 651 798 121 2012
6 NGA Nigeria 6 8.10 212 767 496 9 2012Period Selection Validation:
• Dataset now spans only the years 2012 through 2022 as intended.
• Confirms our analysis will focus on a consistent 11-year window.
• Underlying record count remains unchanged, preserving full coverage within this period.
Identify the eight countries with the highest cumulative fatalities and keep only their records.
top8_countries <- combined_raw_df %>%
group_by(country) %>%
summarise(total_fatalities = sum(fatalities, na.rm=TRUE), .groups="drop") %>%
slice_max(total_fatalities, n=8) %>%
pull(country)
filtered_df <- combined_raw_df %>%
filter(country %in% top8_countries)
top8_countries[1] "Iraq" "Afghanistan" "Nigeria" "Pakistan" "Somalia"
[6] "Syria" "Burkina Faso" "Mali" Top 8 Countries by Total Fatalities (2012–2022):
- Iraq, Afghanistan, Nigeria, Syria, and Pakistan consistently record the highest terrorism-related death tolls each year (IEP GTI map :contentReferenceoaicite:0).
- Somalia remains among the worst-hit due to the persistent al-Shabaab insurgency and large-scale attacks (UN experts report :contentReferenceoaicite:1).
- Burkina Faso and Mali have surged into the top eight as jihadist violence in the Sahel intensifies—this region now accounts for over 40 % of global terrorism fatalities (GTI 2024 Briefing PDF :contentReferenceoaicite:2).
Summarise Annual Fatalities: Aggregate total fatalities by country and year for our selected top-8 countries, convert counts into “thousands,” and print the full 2012–2022 series.
library(dplyr)
annual_fat_df <- filtered_df %>%
group_by(country, year) %>%
summarise(
annual_fatalities = sum(fatalities, na.rm = TRUE),
.groups = "drop"
) %>%
mutate(
fatalities_k = annual_fatalities / 1000
)
# Print all 88 rows to verify
print(annual_fat_df, n = nrow(annual_fat_df))# A tibble: 88 × 4
country year annual_fatalities fatalities_k
<chr> <int> <dbl> <dbl>
1 Afghanistan 2012 1511 1.51
2 Afghanistan 2013 923 0.923
3 Afghanistan 2014 1143 1.14
4 Afghanistan 2015 1008 1.01
5 Afghanistan 2016 643 0.643
6 Afghanistan 2017 1256 1.26
7 Afghanistan 2018 1624 1.62
8 Afghanistan 2019 1194 1.19
9 Afghanistan 2020 1292 1.29
10 Afghanistan 2021 1499 1.50
11 Afghanistan 2022 633 0.633
12 Burkina Faso 2012 0 0
13 Burkina Faso 2013 0 0
14 Burkina Faso 2014 0 0
15 Burkina Faso 2015 4 0.004
16 Burkina Faso 2016 53 0.053
17 Burkina Faso 2017 72 0.072
18 Burkina Faso 2018 144 0.144
19 Burkina Faso 2019 682 0.682
20 Burkina Faso 2020 666 0.666
21 Burkina Faso 2021 759 0.759
22 Burkina Faso 2022 1135 1.14
23 Iraq 2012 2086 2.09
24 Iraq 2013 3980 3.98
25 Iraq 2014 3206 3.21
26 Iraq 2015 2974 2.97
27 Iraq 2016 4514 4.51
28 Iraq 2017 1732 1.73
29 Iraq 2018 1083 1.08
30 Iraq 2019 533 0.533
31 Iraq 2020 464 0.464
32 Iraq 2021 540 0.54
33 Iraq 2022 174 0.174
34 Mali 2012 45 0.045
35 Mali 2013 45 0.045
36 Mali 2014 36 0.036
37 Mali 2015 181 0.181
38 Mali 2016 144 0.144
39 Mali 2017 296 0.296
40 Mali 2018 340 0.34
41 Mali 2019 389 0.389
42 Mali 2020 404 0.404
43 Mali 2021 604 0.604
44 Mali 2022 944 0.944
45 Nigeria 2012 767 0.767
46 Nigeria 2013 931 0.931
47 Nigeria 2014 2101 2.10
48 Nigeria 2015 2003 2.00
49 Nigeria 2016 518 0.518
50 Nigeria 2017 794 0.794
51 Nigeria 2018 642 0.642
52 Nigeria 2019 688 0.688
53 Nigeria 2020 865 0.865
54 Nigeria 2021 497 0.497
55 Nigeria 2022 385 0.385
56 Pakistan 2012 1322 1.32
57 Pakistan 2013 1503 1.50
58 Pakistan 2014 996 0.996
59 Pakistan 2015 658 0.658
60 Pakistan 2016 631 0.631
61 Pakistan 2017 632 0.632
62 Pakistan 2018 502 0.502
63 Pakistan 2019 250 0.25
64 Pakistan 2020 263 0.263
65 Pakistan 2021 292 0.292
66 Pakistan 2022 643 0.643
67 Somalia 2012 446 0.446
68 Somalia 2013 264 0.264
69 Somalia 2014 256 0.256
70 Somalia 2015 335 0.335
71 Somalia 2016 760 0.76
72 Somalia 2017 1535 1.54
73 Somalia 2018 833 0.833
74 Somalia 2019 693 0.693
75 Somalia 2020 671 0.671
76 Somalia 2021 661 0.661
77 Somalia 2022 755 0.755
78 Syria 2012 1014 1.01
79 Syria 2013 907 0.907
80 Syria 2014 545 0.545
81 Syria 2015 226 0.226
82 Syria 2016 531 0.531
83 Syria 2017 376 0.376
84 Syria 2018 323 0.323
85 Syria 2019 368 0.368
86 Syria 2020 746 0.746
87 Syria 2021 494 0.494
88 Syria 2022 447 0.447Fatality Rate Calculation:
• Groups the top-8 countries by country and year.
• Sums fatalities into annual_fatalities (with na.rm = TRUE).
• Drops years with zero incidents to avoid undefined rates.
• Converts to thousands fatalities_k = annual_fatalities / 1000.
• Prints all 88 rows (2012–2022) so you can verify every country–year rate before plotting.
Visualizing Annual Fatalities: A stacked‐area chart showing total yearly fatalities (in thousands) for our top 8 countries over 2012–2022.
library(dplyr)
library(ggplot2)
library(scales)
library(tibble)
library(grid)
# 1) Data prep
df <- filtered_df %>%
group_by(country, year) %>%
summarise(fatalities_k = sum(fatalities, na.rm = TRUE) / 1000, .groups = "drop")
# 2) Stack ordering
order_levels <- df %>%
group_by(country) %>%
summarise(total = sum(fatalities_k), .groups = "drop") %>%
arrange(total) %>%
pull(country)
df$country <- factor(df$country, levels = order_levels)
# 3) Global totals
top_k <- df %>%
group_by(year) %>%
summarise(global_k = sum(fatalities_k), .groups = "drop")
# 4) Anchor labels for 2013, 2016, 2019
annot_years <- top_k %>%
filter(year %in% c(2013, 2016, 2019)) %>%
mutate(label = if_else(
year == 2013,
paste0("Peak: ", round(global_k, 1), " k"),
paste0(round(global_k, 1), " k")
))
# 5) 2020 rebound values
y2020 <- top_k$global_k[top_k$year == 2020]
label20 <- paste0(round(y2020, 1), " k")
# 6) Key events
events <- tribble(
~year, ~date, ~label, ~ypos,
2013, "21 Aug 2013", "Ghouta chemical attack", top_k$global_k[top_k$year == 2013] + 1.5,
2016, "17 Oct 2016", "Battle of Mosul begins", top_k$global_k[top_k$year == 2016] + 1.5,
2019, "27 Oct 2019", "Baghdadi killed by US forces", top_k$global_k[top_k$year == 2019] + 2.5
)
# 7) Palette
okabe_ito <- c(
"Iraq" = "#E69F00",
"Afghanistan" = "#56B4E9",
"Nigeria" = "#009E73",
"Pakistan" = "#F0E442",
"Somalia" = "#0072B2",
"Burkina Faso" = "#D55E00",
"Mali" = "#CC79A7",
"Syria" = "#999999"
)
# 8) Plot
ggplot(df, aes(x = year, y = fatalities_k, fill = country)) +
# a) stacked areas
geom_area(color = "white", size = 0.2, alpha = 0.85) +
# b) global total line (inherit.aes=FALSE so it won't try to map fill=country)
geom_line(data = top_k,
aes(x = year, y = global_k, color = "Global total"),
inherit.aes = FALSE,
size = 1) +
# c) anchor points and labels
geom_point(data = annot_years,
aes(x = year, y = global_k),
inherit.aes = FALSE, color = "red", size = 3) +
geom_text(data = annot_years,
aes(x = year, y = global_k, label = label),
inherit.aes = FALSE,
nudge_y = 0.5, fontface = "bold", size = 4, color = "red") +
# d) downtrend arrow
geom_segment(data = tibble(),
aes(x = 2016,
y = top_k$global_k[top_k$year == 2016] + 0.8,
xend = 2019,
yend = top_k$global_k[top_k$year == 2019] + 0.8),
inherit.aes = FALSE,
arrow = arrow(length = unit(0.25, "cm"), type = "closed"),
color = "red", size = 0.8) +
annotate("text",
x = 2017.5,
y = approx(top_k$year, top_k$global_k, xout = 2017.5)$y + 1.3,
label = "Downtrend",
angle = -23,
fontface = "bold", size = 4, color = "red", hjust = 0.5) +
# e) 2020 rebound vertical
geom_vline(xintercept = 2020, linetype = "dashed", color = "gray30") +
# f) 2020 rebound point, value, and label (each with its own one-row data)
geom_point(data = tibble(year = 2020, y = y2020),
aes(x = year, y = y),
inherit.aes = FALSE, color = "red", size = 3) +
geom_text(data = tibble(year = 2020, y = y2020, label = label20),
aes(x = year, y = y, label = label),
inherit.aes = FALSE,
nudge_y = 0.3, fontface = "bold", size = 4, color = "red") +
geom_label(data = tibble(year = 2020, y = y2020 + 0.9, label = "Affiliate Rebound"),
aes(x = year, y = y, label = label),
inherit.aes = FALSE,
size = 3.5, fontface = "bold",
fill = "white", color = "gray20", label.size = 0.05) +
# g) event vlines and labels
geom_vline(data = events,
aes(xintercept = year),
linetype = "dashed", color = "gray30") +
geom_label(data = events,
aes(x = year, y = ypos, label = paste0(label, "\n", date)),
inherit.aes = FALSE,
size = 4, fontface = "bold",
fill = "white", color = "gray20", label.size = 0.05) +
# h) scales & legend
scale_fill_manual(name = "Country", values = okabe_ito) +
scale_color_manual(NULL,
values = c("Global total" = "black"),
guide = guide_legend(override.aes = list(linetype = 1, size = 1.2))) +
scale_x_continuous(breaks = 2012:2022) +
scale_y_continuous("Fatalities (thousands)",
expand = expansion(mult = c(0.05, 0.15))) +
# i) titles, subtitle, source
labs(
title = "Global Terrorism Fatalities in the Highest-Impact Countries (2012–2022)",
subtitle = "Cumulative deaths peaked in 2013; a pronounced decline followed the Battle of Mosul and ISIS leadership loss, with a modest rebound in 2020",
caption = "Source: Kaggle (Kaggle, 2023): Global Terrorism Index 2023 (2012–2022)",
x = "Year"
) +
# j) theme
theme_minimal(base_size = 14) +
theme(
plot.title = element_text(hjust = 0.5, face = "bold", size = 16),
plot.subtitle = element_text(hjust = 0.5, size = 12),
plot.caption.position = "plot",
plot.caption = element_text(hjust = 1, face = "italic", size = 10),
legend.position = "bottom",
panel.grid.minor = element_blank()
)
Key Insights from Annual Fatalities (2012–2022):
Global Peak & Major Surges: Total terrorism deaths hit their highest point around 2013, driven largely by ISIS’s rise (declared caliphate in June 2014) and the Ghouta chemical attack (Aug 2013). A secondary surge in 2016 coincides with the Battle of Mosul.
Steady Decline Post-2016: After Mosul, overall fatalities fell sharply through 2017–2019, culminating in the killing of ISIS leader Baghdadi in October 2019.
Recent Uptick & Shifts: From 2020 onward fatalities stabilized at lower levels, with a mild rebound in 2021 as the Taliban recaptured Kabul.
Country Contributions: Iraq and Nigeria remain the largest contributors throughout, followed by Afghanistan and Somalia, while the Sahel states (Burkina Faso, Mali) gradually account for a bigger share in later years.
This analysis has examined the Global Terrorism Index data from multiple perspectives, revealing important patterns in how terrorism has impacted different countries over the past decade. The visualizations highlight both the geographic concentration of terrorism and its evolving nature over time.
Several key findings emerge:
Terrorism impact is highly concentrated in a small number of countries, particularly in the Middle East, South Asia, and parts of Africa
The overall global trend shows a peak in terrorism impact around 2013–2016, followed by a general decline
Different countries show distinct patterns of improvement or deterioration in their terrorism situation
The various dimensions of terrorism impact (fatalities, injuries, incidents, hostages) affect countries in different proportions
The severity classification framework provides valuable context for interpreting GTI scores
These insights can inform policy discussions about international security cooperation, humanitarian assistance, and targeted counter-terrorism efforts in the most affected regions.
library(dplyr)
library(plotly)
library(purrr)
library(tibble)
# 1) Data prep
df_plot <- filtered_df %>%
group_by(country, year) %>%
summarise(fatalities_k = sum(fatalities, na.rm = TRUE) / 1000, .groups = "drop")
top_k <- df_plot %>%
group_by(year) %>%
summarise(global_k = sum(fatalities_k), .groups = "drop")
# 2) Palette & stacking order
okabe_ito <- c(
"Iraq" = "#E69F00",
"Afghanistan" = "#56B4E9",
"Nigeria" = "#009E73",
"Pakistan" = "#F0E442",
"Somalia" = "#0072B2",
"Burkina Faso" = "#D55E00",
"Mali" = "#CC79A7",
"Syria" = "#999999"
)
order_levels <- df_plot %>%
group_by(country) %>%
summarise(total = sum(fatalities_k), .groups = "drop") %>%
arrange(total) %>%
pull(country)
# 3) Initialize plotly
p <- plot_ly()
# 4) Vertical dotted guides for each year
for (xv in 2012:2022) {
p <- p %>% add_segments(
x = xv, xend = xv,
y = 0, yend = max(top_k$global_k) * 1.2,
line = list(dash = "dot", color = "gray80", width = 0.5),
showlegend = FALSE, hoverinfo = "none"
)
}
# 5) Stacked areas (no individual hover)
for (ctry in rev(order_levels)) {
sub <- filter(df_plot, country == ctry)
p <- p %>% add_trace(
x = sub$year,
y = sub$fatalities_k,
name = ctry,
type = "scatter",
mode = "none",
stackgroup = "one",
fillcolor = okabe_ito[ctry],
hoverinfo = "none"
)
}
# 6) Global total line — hover turned off here
p <- p %>% add_trace(
data = top_k,
x = ~year,
y = ~global_k,
name = "Global total",
type = "scatter",
mode = "lines",
line = list(color = "black", width = 2),
hoverinfo = "none" # <- disable its own popup
)
# 7) Consolidated tooltip at each year
tooltip_df <- top_k %>%
mutate(
tooltip = map_chr(year, function(y) {
rows <- filter(df_plot, year == y)
# build lines: Year + Global + each country
hdr <- paste0("Year: ", y, "<br>Global total: ", round(global_k[year==y],2), " k deaths")
cntr <- map_chr(order_levels, function(ct) {
val <- rows$fatalities_k[rows$country == ct]
sprintf("<span style='color:%s'>■</span> %s: %.2f k deaths",
okabe_ito[ct], ct, val)
})
paste(c(hdr, cntr), collapse = "<br>")
})
)
p <- p %>% add_trace(
data = tooltip_df,
x = ~year,
y = max(top_k$global_k) * 1.2, # above the stack
type = "scatter",
mode = "markers",
marker = list(size = 0),
hoverinfo = "text",
text = ~tooltip,
hoverlabel = list(align = "left", font = list(size = 10)),
showlegend = FALSE
)
# 8) Final layout
p %>% layout(
title = list(text = "Top 8 Countries: Terrorism Fatalities (2012–2022)", x = 0.5),
xaxis = list(title = "Year", tickmode = "array", tickvals = 2012:2022),
yaxis = list(title = "Fatalities (thousands)", range = c(0, max(top_k$global_k)*1.25)),
legend = list(orientation = "h", x = 0.1, y = -0.2),
margin = list(b = 80),
hovermode = "x unified"
)References
AP News. (2017, December 21). Freedom from IS in Mosul costs lives of 9,000-plus civilians. https://www.apnews.com/article/6d94ba30348e4a25bf76e031b0212ef9
CBS News. (2017, December 20). AP says Mosul ISIS battle civilian death toll 10-times higher than US, Iraq acknowledge. https://www.cbsnews.com/news/ap-mosul-isis-civilian-death-toll-10-times-higher-us-iraq-acknowledge/
CNN. (2016, December 2). Mosul battle: Mounting death toll, urban warfare slow. https://www.cnn.com/2016/12/02/middleeast/iraq-mosul-battle-isis/index.html
CNN. (2016, December 3). Iraq disputes number killed in battle against ISIS. https://edition.cnn.com/2016/12/03/middleeast/iraq-mosul-battle-deaths/
Healy, K. (2018). Data visualization: A practical introduction. Princeton University Press.
Human Rights Watch. (2013, September 10). Attacks on Ghouta: Analysis of alleged use of chemical weapons in Syria. https://www.hrw.org/report/2013/09/10/attacks-ghouta/analysis-alleged-use-chemical-weapons-syria
Institute for Economics and Peace. (2023). Global Terrorism Index 2023: Measuring the impact of terrorism. https://www.economicsandpeace.org/global-terrorism-index/
R Core Team. (2023). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R-project.org/
ReliefWeb. (n.d.). Mosul report 17Oct2016-10Jul2017, FINAL as of 30 October 2017. https://reliefweb.int/attachments/1319cb61-4d76-3cfa-bfd8-c075e91d0801/Mosul_report%2017Oct2016-10Jul201731%20October_2017.pdf
Telegrafi. (2016). Training of Russia’s anti-terrorist troops - like a real event (Video). https://telegrafi.com/en/the-training-of-the-anti-terrorist-troops-of-russias-as-a-real-video-event/
U.S. Department of Defense. (2019, October 27). U.S. forces kill ISIS founder, leader Baghdadi in Syria. https://www.defense.gov/news/news-stories/article/1999751/us-forces-kill-isis-founder-leader-baghdadi-in-syria/
Vox. (2017, December 20). The US spent months battling ISIS in Mosul. At least 9,000 civilians died. https://www.vox.com/world/2017/12/20/16800510/mosul-death-toll-isis-trump-war
Wickham, H., & Grolemund, G. (2017). R for data science: Import, tidy, transform, visualize, and model data. O’Reilly Media.