Εφαρμογή τεχνικών πολυμεταβλητής ανάλυσης σε δεδομένα κινητής τηλεφωνίας
Application of multivariate analysis techniques in mobile phone data
View/ Open
Keywords
Πολυμεταβλητή ανάλυση ; Δεδομένα ; Κινητή τηλεφωνίαAbstract
In the present master thesis, methods of Multivariate Statistics are applied to mobile telephony data. This data was collected by customers from a mobile phone company store in the summer of 2020 and concern subscribers who use a prepaid program from the specific company.
Multivariate statistical analysis is now a valuable tool in the hands of statisticians who are interested in analyzing data from companies. Its methods may be exploited to draw useful conclusions from a large body of data consisting of possibly hundreds of variables, which in turn consist of thousands of observations.
The purpose of this thesis is to present the theory of a number of popular methods of Multivariate Analysis as well as their practical application in real data of a mobile telephony company. Through the application of these methods, the company, by summarizing the information gathered from this large number of observations and variables, can create customer profiles, categorize them into groups, as well as find out what factors influence the subscribers' choices in the market products of the company and specifically their choices in prepaid programs.
The company, based on the results that could be obtained, could define a new strategic plan by determining whether it is necessary to restructure the products it provides, according to the needs of its subscribers or to choose its marketing plan, by analyzing each customer group separately.