Masses in (E)motion

Masses in (E)motion is a moving-image analysis of the 2010 Love Parade disaster, which I developed with Martin Marburger at KISD. It asks whether the density of a crowd, and the moment at which that density tips, can be read from an existing surveillance recording even though no single frame of it shows them.

On 24 July 2010, twenty-one people died of suffocation at the Love Parade in Duisburg, and at least five hundred more were injured. It was the first time the festival had been held in an enclosed area.

For the analysis, the surveillance footage from camera 14 was reduced to one per cent opacity, then duplicated several hundred times and stacked, each copy offset by a single frame, until the sum again reached full opacity. The result is a visual cross-section through a hundred consecutive frames, a moving average of bodies. The sequence was finally accelerated tenfold.

Findings. At first the structural conditions of the tunnel are legible: light from the access ramps still reaches the floor. For roughly the first ninety minutes, visitors stream unimpeded towards the festival grounds. Only when police close access to the main ramp at around 16:00 does the number of people inside the tunnel rise abruptly; the crowd turns into a solid body. The image brightens as a result, since the light is now thrown back by the heads of the tightly packed crowd.

The documentation ends a few minutes before the first deaths on the main ramp. One cause of the disaster probably lies in what this analysis makes visible: those pressing forward had no clear view of the main ramp and could not react to the congestion there, so the pressure on those already standing on it kept increasing from behind.

Techniques: Adobe After Effects · Adobe Illustrator

With: Martin Marburger