Estimating Movement from Mobile Telephony Data


Doyle, John (2014) Estimating Movement from Mobile Telephony Data. PhD thesis, National University of Ireland Maynooth.

[img]
Preview
Download (34MB) | Preview


Share your research

Twitter Facebook LinkedIn GooglePlus Email more...



Add this article to your Mendeley library


Abstract

Mobile enabled devices are ubiquitous in modern society. The information gathered by their normal service operations has become one of the primary data sources used in the understanding of human mobility, social connection and information transfer. This thesis investigates techniques that can extract useful information from anonymised call detail records (CDR). CDR consist of mobile subscriber data related to people in connection with the network operators, the nature of their communication activity (voice, SMS, data, etc.), duration of the activity and starting time of the activity and servicing cell identification numbers of both the sender and the receiver when available. The main contributions of the research are a methodology for distance measurements which enables the identification of mobile subscriber travel paths and a methodology for population density estimation based on significant mobile subscriber regions of interest. In addition, insights are given into how a mobile network operator may use geographically located subscriber data to create new revenue streams and improved network performance. A range of novel algorithms and techniques underpin the development of these methodologies. These include, among others, techniques for CDR feature extraction, data visualisation and CDR data cleansing. The primary data source used in this body of work was the CDR of Meteor, a mobile network operator in the Republic of Ireland. The Meteor network under investigation has just over 1 million customers, which represents approximately a quarter of the country’s 4.6 million inhabitants, and operates using both 2G and 3G cellular telephony technologies. Results show that the steady state vector analysis of modified Markov chain mobility models can return population density estimates comparable to population estimates obtained through a census. Evaluated using a test dataset, results of travel path identification showed that developed distance measurements achieved greater accuracy when classifying the routes CDR journey trajectories took compared to traditional trajectory distance measurements. Results from subscriber segmentation indicate that subscribers who have perceived similar relationships to geographical features can be grouped based on weighted steady state mobility vectors. Overall, this thesis proposes novel algorithms and techniques for the estimation of movement from mobile telephony data addressing practical issues related to sampling, privacy and spatial uncertainty.

Item Type: Thesis (PhD)
Keywords: Estimating Movement; Mobile Telephony Data;
Academic Unit: Faculty of Science and Engineering > Electronic Engineering
Item ID: 5400
Depositing User: IR eTheses
Date Deposited: 16 Sep 2014 11:23
Funders: Science Foundation of Ireland Grant No. (07/SRC/I1168), IRCSET, ERSI
URI:

    Repository Staff Only(login required)

    View Item Item control page

    Document Downloads

    More statistics for this item...