SLIDE 1
Understanding Urban Dynamics with Community Behaviour Modelling
- Dr. Afra Mashhadi
Bell Laboratories Alcatel-Lucent January 2015
Understanding Urban Dynamics with Community Behaviour Modelling - - PowerPoint PPT Presentation
Understanding Urban Dynamics with Community Behaviour Modelling - - PowerPoint PPT Presentation
Understanding Urban Dynamics with Community Behaviour Modelling Understanding our cities and citizens Dr. Afra Mashhadi Bell Laboratories Alcatel-Lucent January 2015 The world is in the midst of an immense population shift 50% living in
SLIDE 2
SLIDE 3
The world is in the midst of an immense population shift
– 50% living in cities; 1.2 billion people to live in urban areas (2050)
How to keep up with changes in the urban cities?
SLIDE 4
Understanding Urban Dynamics with Community Behaviour Modelling
Crowd-sourcing: leveraging the power of citizens to contribute dynamic information about the city.
Directly from the citizens through explicit reporting.
SLIDE 5
Challenge Dataset Results
Can we rely on the quality of the crowd-sourcing to deal with accuracy and coverage issues?
Understanding Urban Crowd-sourcing
Impact
Enabling crowd-sourcing for practical and important problems with massive social impact.
- Accuracy is high and sometimes higher than proprietary data.
- Coverage is low and non-uniformly distributed.
- Coverage is impacted by socio-economic factors.
- Participation is highly affected by the cultural factors.
Given a set of users and profile of an area (e.g., population) we can predict the future crowd-sourcing participation in that area.
Census
6 YEARS OF Points pf Interests (POIs) CROWDSOURCED in 35 COUNTRIES
OpenStreetMap
SLIDE 6
Understanding Urban Dynamics with Community Behaviour Modelling
Crowd-Sensing through infrastructure: Leveraging the citizens footprint through infrastructural sensors. E.g., Oyster cards.
SLIDE 7
Challenge Results
Can we leverage Transportation Data to understand the urban dynamics better, e.g., who, when, (and how often) go to a particular area and why?
Citizen Behaviour Modelling with Transportation Data
Impact
Better Crowd Management, Improved Infrastructure Planning, Adaptive Transport Scheduling. Location Profiling Models with Different Ranking Algorithms for
- Spatio-Temporal Travel Trajectory with Activity Inference.
- Significant Place Detection (Home, Work, Favourite Dining etc.)
- Citizen Behaviour Analysis : Diversity, Introversion, etc.
Given ONLY transportation data, our model can infer the profile of the location and the purpose of the travel.
Dataset
SLIDE 8
Understanding Urban Dynamics with Community Behaviour Modelling
Mobile Communication
Learning more about the urban cities through
- ther sources of data. E.g., Call Data Records.
SLIDE 9
Challenge Dataset Results
Can we leverage mobile communication meta data to understand a community better? Can we model their level of socio-economic?
Understanding Socio-Economic Structure through Communication Patterns
Impact
Radical minimisation of the cost and complexity of traditional methods (household surveys, Census) A set of models based on physics (gravity) and economy (diversity, introversion) that can infer the socio-economic level of areas in fine granularity With meta communication data we can infer the state and change of socio-economic level
- f a community in fine granularity that was not possible before.
Global-Pulse UN research lab
1 YEAR aggregated CDR data Ivory coast