TU Berlin

Service-centric NetworkingLüders, B. (2016). Macroscopic Pattern Mining based on Human Mobility in Urban Environments, Master Thesis. Technische Universität Berlin


zur Navigation

Es gibt keine deutsche Übersetzung dieser Webseite.

Master Thesis: Macroscopic Pattern Mining based on Human Mobility in Urban Environments


Macroscopic Pattern Mining based on Human Mobility in Urban Environments


Nowadays, mobile devices are ubiquitous and nearly seamlessly integrated in daily life. They allow to make use of sensor data, capturing the surrounding context, which led to the emergence of a broad range of context-based services. Many individuals rely on these, commonly referred to as mobile applications, to support them in their daily activities. A major branch of these services are location-based services, which utilize the location information of the mobile device to deliver appropriate, location-related content in a variety of applications, ranging from navigation and route tracking to location-based reminders and location recommender systems. In recent years, these location-based services changed their functionality from being mainly reactive, delivering content based on an individual’s request, to proactive. The proactiveness is enabled by robust positioning technologies and the development of background tracking and geofencing. These two major technological advances allow to monitor the user’s location in the background and to proactively support the user. This process transforms the physical, real-world movement into a representation in terms of data. Due to the rising popularity of location-based services, this leads to a constantly increasing asset of mobility data, which can be used to support urban planning and transport, to understand commuter flows, to gain insights into touristic behaviour and finally to develop new services and offers. Considering the steadily increasing population in cities, urban planning is currently one of the most significant application scenarios and can be significantly supported with location and mobility analytics. On this basis, the thesis presents a new approach to comprehend macroscopic mobility behaviour based on coarse mobility data. The thesis provides a conceptual framework and demonstrates the prototypical implementation of an algorithm to cluster individuals, based on their macroscopic mobility behaviour, into groups. Building on an extensive review of related approaches, statistical methods to process mobility data and a summary of research in the area of human mobility patterns, concepts for the extraction of features to capture and measure macroscopic mobility behaviour are developed and implemented. The evaluation of the results illustrates that a set of these features allows to precisely characterise individuals, without revealing their identity or actual whereabouts, and to generate universal mobility prototypes. These, as well as the intermediary results, yield valuable insights into the variety of human mobility behaviour.

Supervisor: Axel Küpper, Peter Ruppel

Type:  Master Thesis

Duration: 6 months



Schnellnavigation zur Seite über Nummerneingabe