Ph.D. Student Kai Zhao from our department together with three others students from Aalto University (Nautiyal Sudhansu, Eranti Pradeep, and Tatiraju Venkata) won one of the best hack awards in the Data Science Hackathon competition. They analyze the HSL dataset (Helsinki Open Bus Dataset provided by HSL) with the statistical and network analysis methods and they have some interesting findings.
First they observe that during peak time most of the Most of the bus delay (Helsinki City Traffic) happen during a certain Hot Areas such as Central railway station, Pasila station, the key entrance road to the city center and the ring roads near airport (see Figure 1). They identify these key areas using the Betweenness Centrality and plot them in Figure 2. They use the Helsinki bus delay as an identification of the urban traffic for analyzing the Helsinki City Traffic.
They also find that in Helsinki the Bus traffic is not an important cause of the City Traffic. They use the Pearson coefficient to verify the correlation between the bus traffic and city traffic and they are not correlated with a Pearson value of 0.013. To reduce the Helsinki traffic, they suggest that we should add more buses during the peak time for reducing other vehicle usage and reducing passenger picking-dropping time.
Figure 1. Helsinki Traffic on Monday morning between 8 am and 9 am (2014-02-03). The darker, the longer delay of the buses (the more city traffic) on that road.
Figure 2. Key areas causing most of the traffic in Helsinki during peak time.
Figure 3. Statistical results show that there is no correlation between bus traffic and Helsinki urban total traffic. It means that the buses are not an important cause of the Helsinki traffic, other vehicles are.
Kai Zhao's homepage: https://sites.google.com/site/zhaokaics/
Results: http://datasciencehackathon.cs.hut.fi/?page_id=115
Slides: http://datasciencehackathon.cs.hut.fi/wp-content/uploads/2015/04/PVK.pdf