(Big) Data Analysis for Mobile Operator

Supervisor: Martin Kruliš
Intended Scope: Master thesis or software project
Required Skills: data mining and data analysis, other required skills will be specified based on selected project


This project will specified in cooperation with T-Systems Competence and Delivery Center Prague, which is a research lab of T-mobile (Deutsche Telekom) for big data. The mobile operator has huge amounts of data regarding cell phone usage (call records, data traces, billing, location of the devices within the cellular network etc.). These data may be used for data analytics to gather important information, which can be either used by the operator to improve services or marketing, or to be sold to other parties.

There are several topics, in which the CDCP is currently very interested:

  • Location insights — i.e., analysing location and movement of the people to generate heatmaps (where people of particular demographic groups gather or travel), performing urbanistic studies, or traffic studies (e.g., traffic jam predictions).
  • Household detection is identification of interesting locations and social groups for each individual. This means detecting locations such as home, workplace, sports club etc. including identifying related social groups (family, work colleagues, sparring partners, ...) based on repeating user activities.
  • Clustering, similarity/anomaly detection of the users. In other words, clustering algorithms will be applied on given criteria to detect similarity (typical behaviour of customers), anomalies (potential frauds), and some forms of smart user categorization. Furthermore, this analysis should be performed in runtime on streaming data.

The final specification of software project or Master thesis will be concretized based on a discussion between the student(s), supervisor, and CDCP representative, which will also act as a consultant. Please note that this is not a typical SW project as it involves a great deal of research and utilization of existing infrastructures. On the other hand it is expected that SW prototype (proof of concept) will be implemented.

For more information see Promo PDF.

The content is available under Creative Commons BY-NC 3.0 License