Tuesday 2:45 p.m.–3:25 p.m.
Using Massive Passive Phone Calling Data to Understand and Predict Human Mobility
Fahad Alhasoun
- Audience level:
- Intermediate
Description
The talk is centered around using massive phone calling data to capture interesting patterns of how people move in cities. The first part of the talk will show how we can employ simple statistical and visualization tools to unveil interesting mobility patterns. The talk will extend to predicting people’s next visited locations using an implementation of a dynamic bayesian network framework.
Abstract
With the rapid adoption of pervasive technologies, a significant portion of the worlds population utilizes mobile phones where researchers today are using data generated from such technologies to better understand human behavior at unprecedented scales. Such data driven research unveiled statistical patterns that provide understanding of how people communicate, feel, move and so forth.
This talk will be providing the tools and approaches to better understand how the structure of a city derives how humans move, chose a place to live/work and form communities that also in turn influences and derive business opportunities. On the second part of the talk, we take a look at state of the art human mobility prediction algorithms using phone calling activity as a proxy to their locations.
The talk involves using several python libraries such as scikit-learn, numpy and scipy in addition to the Hidden Markov Model (HMM) toolbox on matlab. The goal of the talk is to give the audience a general perspective of the approaches and tools used with mobile phone calling data.
