Mobile phone usage during driving causes driver distraction, which negatively affects road traffic safety and is becoming an increasingly common cause of traffic accidents. This paper proposes a new methodology, based on anonymized datasets from telecommunication networks and analytical possibilities of data science, which has as its goal to identify a pattern of usage of the mobile phone (in particular voice calls and applications) by drivers while driving. Based on this finding, the methodology is used to identify segments of the road infrastructure on which mobile phone usage is more intense. Applying this methodology enables stakeholders in the road traffic safety system to propose targeted measures (educational, legal, or engineering) precisely on the sections where it is most vital to do so in order to reduce mobile phone usage and enhance road safety. The proposed methodology was implemented and demonstrated on two pilot sections on state roads in the Republic of Croatia. Research results point to the fact that the percentage of drivers who use the mobile phone for voice calls or for app usage is exceptionally high (10.68% to 33.90%) and presents a serious threat to road traffic safety.