Mobility Profiler

Mobility Profiler: A Framework For Discovering Mobile User Profiles

In this project, we focus on the problem of discovering mobility patterns and mobility profiles for mobile users from the cellphone-based location logs. We define the mobility path concept used for representing users travel path from one location to another location. We construct the mobility paths of cell phone users from cell tower connection data. We also propose a formal model for mobile user profiles that integrates both frequent mobility patterns and contextual time data for the user. We also present a complete framework, Mobility Profiler, for discovering mobile user profiles starting from cell based raw location data.

Figure 1: Mobility Profiler Framework

The General architectire of Mobility Profiler Framework is given in the figure above. In this framework, we start with path construction which is an essential task for grouping related cell tower connection transactions together. The main goal of this process is constructing an ordered set of cell tower ids correspond to user’s travel from one location to another. After this task, topology construction process is executed. In this process, cell tower neighborhood topology is extracted from mobile user paths obtained in the first step. The topology information is used for eliminating majority of candidate path sequences to expedite pattern discovery phase. This step also provides visualization of cell tower neighborhood information. Pattern discovery is the next step in our framework; it is the major step which discovers core knowledge from path data. There are two main dimensions of pattern discovery step which are discovering global and personal mobility patterns. In global pattern discovery, we analyze the common mobility patters shared among all users. In personal one, we discover frequent mobility pattern for each user separately. Both of these pattern discovery tasks are executed efficiently by using topology information and string matching support criteria. In post processing task, we generate cell phone user profiles from personal mobility patterns by adding time contextual data. In this way, we obtain a powerful representation of mobile user profiles containing both time and location data. Possible Applications of Our Framework are given below:

Mobile User Profiling (pre)

Air Pollution Estimation (pre)

Cell Tower Localization and Oscillation Elimination over Cell Span Data

Using Mobility Behaviour for Opportunistic Routing Algorithms

Social Network Analysis

Project Members

Murat Demirbas, Principal Investigator, Asistant Prof. at SUNY at Buffalo

Murat Ali Bayir, Co-Investigator, PhD Student at SUNY at Buffalo

Reseach Collaborators

Nathan Eagle, Reseach Scientist, MIT

Carole Rudra, Asistant Prof. at SUNY at Buffalo

Atri Rudra, Asistant Prof. at SUNY at Buffalo

Publications

Murat Ali Bayir, Murat Demirbas, Nathan Eagle, Mobility Profiler: A Framework for Discovering Mobile User Profiles, 2008 (Under Submission)

Murat Demirbas, Carole Rudra, Atri Rudra, Murat Ali Bayir: IMAP: An Indirect Measurement of Air Pollution via Cell Phone, 2008 (Under Submission)

Navigation

News