From big data to smart data, traffic optimization using mobile network traces

Swisscom mobile network generates 20 billion events per day (connection of a mobile phone to an antenna), equal to 2 Terabytes of data.

What if these anonymized and aggregated mobile network traces could help cities make the right decision for their infrastructure investment, reduce your time in transport, help you shift to greener mode of transport and ultimately improve quality of life and reduce CO2 emission in the city.

This blog post will summarize what is the opportunity, what are the applications, where we stand and where we are aiming to.

The opportunity

Addressing mobility challenges required first to have a good understanding of the traffic situation, i.e. how many people are in town, from where do they come from, where do they go, …

To achieve it, our vision is very simple: we want to build the analytic tool of the physical world

Tool

We have plenty of tools available to monitor the traffic in the Internet world and now technology enable us to develop the tool for the physical world.

Swisscom is well positioned to innovate in that field since we have a high representativeness (mobile phone market share of 60%), good 4G coverage (97% of the population) and high penetration of smart phones.

In traditional traffic survey, you capture an image of the traffic for a given day or a week. We are now shifting from a single snapshot to a complete movie, as illustrated with the video below visualizing the traffic for the city of Zurich during one complete day.

The applications

There are obviously plenty of applications, I will hereby describe 3 concrete ones aiming to improve the traffic by setting up the right infrastructure:

Construction of a new road section to reduce transit traffic: when investing 50 MCHF in a new road to offload the transit traffic downtown, our tool can first justify the rational for this substantial investment and measure the impact before and after the implementation.
Expansion and transformation of a parking: before investing 10 MCHF in a parking, our tool can help identify the user needs. For example, the setup might be different if the parking is used by drivers for carpooling of if it is used to drop the car and then use public transport to go to work.
Deployment of charging stations for electrical cars: we can help defining the best location to install charging stations and therefore make it more convenient to shift to electric cars.

The inital results

We have started working on a co-creation mode with the City of Pully and the Canton of Geneva. The three initial mobility insights results are:

We measured the transit traffic in the center of Pully and its origin: we observed 25’000 trips going through Pully during a day and 75% of this traffic is transit i.e. people spending less than 15 minutes in the city. In addition, we also visualized below the heatmap of the origin of this flow of traffic.
We observed the “lignes de désir” in the Canton of Geneva: we were able to define the most used roads. We visualized the changes during the day and how it developed over a week.
We created a heatmap showing the regional distribution of visitors coming to Geneva International Motor Show: we can therefore answer a key question of event organizers: from where are my visitors coming from ?

The future

As you can see for the above, we are just at the beginning of an exciting development. The first step is to build this unique tool to measure continuously the traffic flow in the physical world. Then in the future, we will be able to enrich it with additional data sources like traffic light information, public transport data, car sharing solutions, building energy consumptions, … to capture the full potential of a Smart City i.e. leverage all the data available and convert them into “smart data” to run the city in the most efficient way to better serve their citizens.

The initial results with the City of Pully were presented at EPFL during the Connected Event, “from big data to smart data” which took place on the 7^th of October 2015. Upcoming results will be presented at the 4^th National Smart City conference taking place in Zug on the 3^rd of December 2015.