Analyzing people’s mobility patterns for the purpose of the urban public transport network optimization. We developed the methodology to analyze origin-destination matrices provided by the cell phone operator, in order to highlight areas of residents' concentrations, interconnected by steady mobility flows. Also, public transport network optimality criteria were determined: share of population covered, total routes length, average travel time, and the average number of transfers. Given these areas and key mobility flows, we modeled new optimal public transport network routes, allowing us to achieve the maximum share of trips served with a minimum total length of all routes.