Data Visualization

Finally I have reduced the size of my IABSE 2007 paper within 8 pages. Hopefully this will be my first paper published on a influential conference and a good start for my academic career.

So what is today's new idea? Well, yesterday I read some slides on the TRB web page. That presentation is talking about highway safety. One thing impressed me is that they are trying to visualize the safety status of highways: red means dangerous, white means pretty safe. I do not know any details about how they define the safety level of a section of highway, but I find that data visualization is such a simple and effective way for data interpretation and information management.

We can generate a map labeled by the accident rate, and another map labeled with the population density, the curvature of the road etc. Then we can compute the correlation of accidents and various factors using these labeled maps. In my mind, those maps are images and can be processed by any image processing techniques and pattern classification algorithms for data interpretation. I do not know whether this is data mining and knowledge discovery, but I think image processing will be a powerful tool for us to digest huge amount of data and find patterns and knowledge hiding behind them.

Accident map can have different views. We can classify the accidents into several categories according to a specified taxonomy, and we can even compare different taxonomy based on the generated image and its processing results: a better taxonomy will generate images which can explain existing phenomenons and predict some accident correctly. We can also have hierarchy of the visualization: high level taxonomy can generate high-level images shows high-level trends and sub-taxonomy can zoom in to a specific level and view the data in a specific perspective.

These maps will be very useful for accident reducing since people can use it to guide better transportation planning, drivers can use it to generate a "safer path" based on the start point and the target information. Transportation agencies can use it to manage transportation system maintenance and replacement, and they can also provide guidance to drivers so that keep the traffic on deteriorated highway at a low level.

Visualization, such a good idea!

The link of that presentation is:
http://www.trb.org/conferences/e-session/2007am.htm#DATA