A first glance at the implementation of our designs

In this post we will discuss our progress on implementing the designs from the previous blog post. When exploring the design space with pen and paper, there are no limits to creativity. Once designs shall be implemented with software, challenges can arise. Working with visualization tools like Vega, Vega-lite, p5 and D3 is demanding and can be frustrating at times. We tried to implement three different designs.

Implementation 1

Bubble plot of Al-Qaeda’s attack-types and caused deaths per year

For the first implementation we used the Vega online editor. We cleaned the dataset, used a pivot table to get the relevant information and uploaded the dataset to a Github account. From there, the dataset could be accessed by the Vega editor. Next, a predefined schema was adapted to fit our needs. The code can be found here. The plot displays three dimensions without overwhelming the viewer. It starts in the year 1992, where an explosion carried out by Al-Qaeda killed 2 people. One can clearly see the attacks of 9/11 (Hijacking of planes) as the biggest triumph of the organization. Bombings and explosions are causing a significant amount of deaths every year.

Implementation 2

Sankey plot of attacks’ typology (perpetrator, target and weapon)

For our second implementation, we attempted to reproduce the alluvial diagram that we initially designed to describe the typology of attacks across the main organization and its affiliates. In particular, in addition to the perpetrator’s name, we considered the target and the weapon used in the attack. Given that some groups conducted a relatively small number of attacks, it was hard to identify some of them in the alluvial diagram (since bars are usually stacked). For that reason, we switched to a sankey plot which, although similar, allows for a more sparse disposition of bars representing categories. The diagram above shows from left to right the connections between perpetrators, targets and weapons used. The image is just a snapshot of the plot but in reality the diagram allows for interactively accessing the number of attacks perpetrated of a given kind. For example, it can be observed in the plot that Al-Shabaab (Al-Qaeda’s affiliate in Somalia) has carried out 1,806 attacks to military facilities. As for further developments of this plot, we are considering the inclusion of other variables and/or a temporal tooltip.

Implementation 3

Ridgeline plot of Terrorist Group Activities

For our third implementation, we sought to answer the question of how terrorism activities have evolved for the various Al-Qaeda sub-groups. A Ridgeline Plot was used to represent these dynamics. A Ridgeline Plot is formed by partly overlapping line plots, presenting the impression of a mountain range, or ridges, as seen in the visualization above. The years 1992 to 2019 are given on the x-axis, which are the years in which at least one terrorist group was active. The height of the line represents the proportion of terrorist attacks carried out by a given terrorist organization out of all terrorist activities carried out by that group in a given year. The peak of each line, therefore, represents the year in which a particular terrorist organization was most active. This visualization helps the observer to trace how terrorist organizations’ activities have evolved over the corresponding years, as well as see when specific groups were most active and understanding the dynamics between the groups.