Good question.

I need to imagine:

If such a big city as at least one airport, the real-time data of arrivals can be used to schedule trains and buses to pick up passengers. Those trains are there, when they are needed. Compare this to a classical scheduling by some time slots (such as rush hours etc): if a plane is late, the schedule does not make sense anymore.

In fact, the management of the ground crew for airlines has been optimized by big data analysis and real time information, so it would be an enhancement of that application.

Second idea: using bug data for planing of workforce streams inside the city itself. Who is going to which location at which time? I could imagine, that this level of detail could improve not only the scheduling of public transportation but also planning of buildings (where) and work (when). It could also integrate private transport (by cars) in allowing to find best routes (using big data, we could try to optimize to a maximum) as well as in optimizing sharing cars.

This is a sensor net issue

a) you want sensors to measure traffic density (not hard to do)

b) you want individual sensors (GPS trackers) on cars so you can

determine families of trajectories (this is hard)

Note this problem is highly non-linear - every attempt to control the current state of traffic will immediately send it in to a new state - too many vehicles and too little infrastructure.

Before problems such as congestion issues can be solved with big-data, there are many inherent challenges in big-data that one need to address.

1. Real-time data transfer and aggregation is still in infancy today with big-data.

2. Swarm intelligence is not really in a stable state to be usable by industry. Most of the attempts are just academic proposals.

3. Vehicle-to-Vehicle (V2V, M2M) communications need reliable, low-energy wireless communication protocols, that are still not available (MQTT is still in-working phase). 4G-LTE is still not widely accepted.

4. Sensor devices are not generic enough today. We are still at the level of smart-phones, not smart-devices. No single coherent OS or proposal present for smart-devices. Raspberriens are actively working on this, but still in infancy.

All the above are the basic necessities before big-data can become usable in real-life. Active research is going on all these fields, soon we may have reliable solutions.

GK

http://gk.palem.in/

Article M2M Telematics & Predictive Analytics

For some years now, anonymous gsm tracking is used in Belgium, which make it possible to follow the progression of the vehicles in real time. This is for instance used by radio stations to announce the current time needed on different strategic road chunks. 

OpenStreetMap is an excellent data source for computing and analysis. Based on the analysis, one can conduct simulations to figure out the underlying mechanisms of transportation, how does it work?

Jiang B. and Liu X. (2012), Scaling of geographic space from the perspective of city and field blocks and using volunteered geographic information, International Journal of Geographical Information Science, 26(2), 215-229. Reprinted in Akerkar R. (2013, editor), Big Data Computing, Taylor & Francis: London.

http://fromto.hig.se/~bjg/movingbehavior/

http://fromto.hig.se/~bjg/axwoman/