BUGBusters — Isabella Tagliafico, Luca Ninivaggi, Flavio Barrara Stefani
The BUGBusters study group is composed of three students enrolled in the Master’s Degree in Artificial Intelligence at the University of Genoa, within the Data Visualization course taught by Professor Annalisa Barla.
The goal of the project is to provide an intuitive and informative visualization of a dataset on global terrorism, revealing significant trends, patterns, and correlations. The work aims to turn complex data into clear, meaningful insights that promote a deeper understanding of global security phenomena.
The Global Terrorism Database (GTD) is an event-level dataset maintained by the University of Maryland. It contains over 200,000 records of terrorist attacks worldwide, spanning from 1970 to 2020.
Data are collected from multiple open sources. The GTD research team integrates and standardizes them to ensure consistency over time, performing rigorous quality control, retrospective coding, and cross-referencing with other terrorism databases to identify and include missing events.
Each record in the GTD represents a single terrorist incident and includes detailed information such as the date and location of the attack, type of weapon used, targets and victims, perpetrators, and any claims of responsibility.
Additional variables describe casualties, outcomes, and data sources. A full codebook and documentation can be found on the GTD website.
This section compares key facets of terrorist activity across countries, regions, attack types, targets,
and groups.
Each chart provides a different analytical lens (from geographic concentration to tactical composition and
casualty severity) to help identify patterns worth deeper investigation.
This Bar chart highlights the ten countries with the highest number of recorded terrorist
attacks.
Hover a bar to see the three most affected cities within that country (local concentration vs
national spread).
The following Stacked Bar chart compares the percentage composition of attack types across
regions.
Regional differences suggest distinct operational contexts and strategic goals.
Waffle chart (Icon-based) proportional view of global attack types; each cell represents a share of total incidents. Helps quickly grasp the mix of tactics at a glance.
This Heatmap uses color intensity to indicate incident counts by perpetrator group and victim type (military, police, civilians, businesses…). Useful to identify groups’ preferred targets and potential escalation patterns.
For the ten most active groups, the Multiple Bar chart compares attacks with zero victims, one
victim, and multiple (>1) victims.
Click the legend to isolate a category for focused analysis.
Temporal and Quantitative Dynamics of Terrorism explores how terrorist activity evolves across time, regions, and groups. While the previous section focuses on categorical comparisons, this section shifts the perspective toward distributions, trends, and variability.
The Mirror Plot contrasts the yearly number of fatalities (above the axis) and injuries (below the axis). This mirrored perspective should emphasise asymmetry between deadly and non-deadly consequences of attacks, helping reveal years where harm increased despite fewer deaths, or vice versa; in our case data seems to be symmetric instead.
Data missing in the year 1993
The Ridgeline Plot visualizes how attack intensity varies across regions over time. Each ridge represents one region, displaying its annual attack distribution as a continuous or discrete curve.
The Box Plot visualizes how long attacks last for the most active terrorist groups, showing the distribution of event duration in days. It highlights which groups tend to conduct short, contained incidents versus those associated with prolonged or multi-day operations. The interactive zoom allows a closer inspection of each group’s temporal footprint.
This section presents a time-based narrative of global terrorist activity. Through a timeline, you can visually explore when key incidents occurred.
This Connected Scatter plot visualizes the evolving relationship between the number of terrorist attacks and the total number of victims across years. Each point corresponds to a year, linked chronologically to show the trajectory of global terrorism. Key years are annotated to highlight historically significant shifts—such as sudden increases in attack frequency or dramatic surges in lethality. The visualization helps reveal whether terrorism has become more frequent, more deadly, or both, and how these patterns changed in response to major geopolitical or organizational events.
Terrorist activity is deeply shaped by geography. This section visualizes how attacks are distributed across the world, highlighting both global patterns and country-level concentrations. Through spatial aggregation and proportional symbols, it becomes possible to immediately compare the intensity of terrorism across continents, identify regional clusters, and observe the influence of major groups in specific territories.
This proportional symbol map displays the global distribution of terrorist attacks by placing circles over each affected country. The size of each bubble is proportional to the number of incidents. The color indicates the perpetrator group, highlighting the presence and reach of the ten most active organizations worldwide, while all others are unified under a neutral grey category.
Visualization of geocoded deadly events. Each circle is placed at an event’s coordinates, sized by the number of reported fatalities. Use the timeline controls to explore how the spatial burden of violence evolved over time.
A discretized choropleth visualization illustrating total fatalities at the governorate level. The legend shows numerical thresholds for each color class.
This section examines how terrorist attacks are related in categories, like weapons and targets, to reveal structural patterns behind operational choices.
This Sankey diagram shows how different attack targets flow into specific type categories.
This radial network visualizes primary weapon types and their operational subcategories. Node size reflects frequency, revealing dominant weapon families and their most used subtypes.