Dynamic Networks

One of my research interests is algorithms and data structures for dynamic graphs and networks. In the context of the LIVE Singapore project, I worked on efficient algorithms for large-scale transportation problems.

I had the pleasure and the opportunity to work with Xiaoji Chen, a very talented and creative designer. She made some beautiful visualizations (see video below) of a dynamic network we obtained by combining GSP traces from tens of thousands of taxis. An interactive map was showcased from April to May 2011 in an exhibition at the Singapore Art Museum (see also a similar visualization of public transportation in France).

Watch how the map of Singapore changes throughout the day. As distances in the visualization correspond to average travel times in the city, we can observe phenomena such as the impact of the «rush hour» on commute times. Note that, since the travel time metric is not necessarily symmetric, the map may show a different behavior for different starting points. The video below illustrates this asymmetry by showing a dynamic map for two different starting points, one being in downtown and the other one being at the airport.


If the video does not work, you may download it here (mp4, 21.8 MB).