Final qualifying work of Ermolaev Alexei Nikolaevich devoted to the popular direction of research in the field of data extraction from the Web, in particular, data on participants of social networks for the purpose of using them in models of web graphs for the analysis of the behavior of participants. As a real network has been selected the popular social networking service VKontakte. Such a task is of great practical importance in the era of global changes and will undoubtedly be useful not only in assessing its own audience, but also to identify various kinds of dependencies in public relations. To solve the problem Ermolaev A.N. it was suggested to solve the following subtasks: 1. To develop a program for extracting and systematizing semi-structured data characterizing network participants. 2. To develop a method for constructing graphs of a special kind, which will be discussed below. 3. Develop a program to calculate some metrics of a social graph. 4. Conduct a series of experiments on real networks, analyze and summarize the results. A good analysis of the literature on the study, including articles on social networks, general concepts, history of development, methods of analysis of social networks and their mathematical modeling, the specific possibilities of extracting information from Vkontakte, metrics, assorted, mixing and clicks in graphs, additional features of the Gephi system and the programming language Python. The concepts and methods of data collection for the construction of so-called "small" communities were defined: - ego-count friends; - count friends and their friends; - graph of community users. For the study of small communities, the program SGMaker was developed with a graphical interface, which is of practical value for researchers in the field of social networks, greatly simplifying and speeding up the process of collecting, processing data, initial processing and their preparation for analysis by more powerful software tools. The following software tools were used in its development: Python, NetworkX, VK API, vk_requsts and PyQt, as well as Gephi program for further processing of the results. A series of experiments: 160 different ego-graphs for which the statistical regularities of behavior, ego-ego, count (about 35,000 participants), very close to the graph Barabasi-Albert graphs user community, characterized by weak connectivity. All the results and the developed program open a wide range of tasks for researchers. Manager would like to continue working with Ermolaev A.N. in case of admission to graduate school. Moreover, a joint article has been prepared for publication and submitted to the editor. The final qualifying work of the bachelor Ermolaev Alexey Nikolaevich "Development of methods and algorithms of social networks analysis with the use of theory-graph approach" deserves the evaluation of "excellent".  Scientific adviser, Dr. of Comp. Sci. A.A. Pechnikov