MOBILITAPP: ANALYSING MOBILITY DATA OF CITIZENS INTHE METROPOLITAN AREA OF BARCELONA.

Silvia Puglisi, Ángel Torres Moreira, Gerard Marrugat Torregrosa, Mónica Aguilar Igartua, and JordiForné

Department of Telematics Engineering Universitat Politecnica de Catalunya

GOODTECHS 2015 - Rome, Italy

AGENDABackgroundMobility Master PlanPlatform architectureMobile applicationMobility patterns recognitionPrivacy conscious analyticsConclusions and future work

BACKGROUNDMobility and transportation efficiency have always beenessential in a city for it to function properly.

Smart mobility solutions provide efficient, safe andcomfortable transport services for isitors and residentsalike.

Smart mobility services are information driven and rely ontechnology to provide personalised services to its users.

MOBILITY MASTER PLANThe flow of people and goods over the transport network ofa city is a complex problem.

Transport service providers need to forecast demandaround the city and plan long term investment.

Data about the mobility patterns of users in metropolitanareas is aggregated from partial ticket sales data, surveysand, in some cases, economical models.

METROPOLITAN REGION OF BARCELONA

METROPOLITAN REGION OF BARCELONA:

3,240 Km25.0 million people

CITY OF BARCELONA:

100 Km21.6 million people

The Metropolitan Region of Barcelona is divided into 582zones.

PLATFORM ARCHITECTUREMobilitApp is a smart city platform to collect and analysemobility data of the citizens in the metropolitan area ofBarcelona, Spain.

Mobilitapp synchronously collects updated geographicalposition as well as users’ activities and events:

Traffic information in real-timeTraffic incidencesUser mobility patterns.

Users’ activities are collected considering different sourcesand sensors data. This information is processed to obtain

citizens’ mobility patterns.

Aggregated users data can then be filtered and analised bytransportation authorities and service providers.

Users can access integrated real-time traffic stateinformation provided by the Barcelona City Hall and the

traffic incidence information provided by the Spanish TrafficAuthority: Dirección General de Tráfico.

MOBILE APPLICATIONMobilitApp collects citizens’ mobility patterns in thebackground.

MobilitApp uses Google Android APIs to discover the user’spositions.

Google Android APIs provide a low consumption mechanismto log periodic updates and detected activities.

The APIs use mainly the device accelerometer.

We call the Activity APIs to sample the obtained results every20 seconds.

Every 2 minutes, the algorithm makes a statisticalestimation of the most probable result out of the lastsamples.

We also consider the following factors

Accuracy of the GPS: when a device is underground, theGPS accuracy de- creases consistently.Location of points of interest(POI) to help the algorithmknowing if a user is close to a bus stop or a metro station.Directions: we use Google Directions APIs to check if thereis a known route (using all possible transportation types)between two points.

MOBILITY PATTERNS RECOGNITIONMobilitApp is able to successfully classify between thefollowing activities:on foot: Activity type returned if the citizen is eitherwalking or running.bicycle: Activity type returned if the citizen is on a bicycle.vehicle: Activity type returned if the citizen is on a motorvehicle (e.g. car, motorbike, bus,...).still: Activity type returned if the citizen is not moving.unknown: Activity type returned if Activity RecognitionAPI is not capable to estimate the actual activity.

A key challenge was to successfully classify different types ofvehicles and distinguish between private and publictransportation.

We use a simplified geofencing technique to identify if theuser is using public or private transportation.

For example if we observe that a user has lost GPS contactwhile moving we might assume they have used the metro.

PRIVACY CONSCIOUS ANALYTICSMobilitapp collects a vast range of possible sensible userdata

To avoid exposing users to direct threats of collection andprocessing of private information, MobilitApp has the optionnot to supply any personal details to the platform.

Furthermore, user data collected by our mobile applicationis communicated encrypted to the server.

CONCLUSIONS AND FUTURE WORKWe are determined to continue developing MobilitApp andimprove how we de- tect the user transportation mode andposition.

We are considering to implement methods for locationdiscovery without the use of GPS, to reduce device batteryconsumtion.

We are continuing researching measures to reduce theusers’ privacy risk of sharing sensible data to our platform.

THANK YOU."What is the city but the people?"- William Shakespeare, The Tragedy of Coriolanus