Inhalt des Dokuments
Es gibt keine deutsche Übersetzung dieser Webseite.
Master Thesis: Utilizing Call Detail Records for Travel Mode Discovery in Urban Areas
Call Detail Records for Travel Mode Discovery in Urban
Mobile network operators often bill their customers based on their network usage. For this purpose, operators collect information about billable events, such as calls, text messages, and data usage. In recent years, operators have realized that they can monetize these billing records by selling insights extracted from them. In this thesis, a multi-stage data analysis algorithm is presented that uses these billing records for travel mode classification. This algorithm identifies whether a mobile phone user has traveled using a public transportation bus or using another transportation mode. The billing records collected by a network operator contain the time at which a billable event happened, as well as the network cell from which the event originated. The coverage area of each network cell is known to the operator. Therefore, the billing records of a mobile phone user give an overview of that user’s approximate location at different times. This data can be used to discover the sequence of network cells that the user has traveled through during a trip. Travel mode classification algorithms in literature analyze long-distance or medium-distance trips. The data analysis algorithm presented in this thesis is novel for analyzing and classifying short-distance, intra-city trips. To classify mobility traces, it uses publicly available bus timetable data and road network infrastructure data. The accuracy of the classification algorithm is evaluated using a two-fold cross-validation analysis.
Supervisor: Sandro Rodriguez Garzon 
Type: Master Thesis
Duration: 4 months
10587 Berlin, Germany
Phone: +49 30 8353 58811
Fax: +49 30 8353 58409