Aplikace pro vizualizaci vývoje komunit v sociálních sítích

Abstract

This master's thesis focuses on the design and development of an application for visualizing the evolution of communities on social networks. The application will support interactive and experimental work aimed at monitoring the development of communities. The newly created system will be documented and described in detail. A review of systems addressing similar issues will be carried out to provide an idea of similar types of applications. Experiments on small networks with roughly hundreds of vertices are also an integral part of the work. Each experiment must contain a description and subsequent evaluation. For a detailed understanding of the entire context of this work, a description of graphs, networks, communities, and algorithms, such as the calculation of ego zones, must also be provided.

Within the scope of this work, a web application was created using the Angular framework, which is based on a component architecture with services and supports the TypeScript language. In collaboration with several freely available libraries, the application offers comprehensive options for analyzing temporal networks. The system supports the import of various data sets, the configuration of calculations, and visualization, in which all available objects are displayed. They can be moved, searched, and colored using algorithms. Basic work with a view of the diagram, such as zooming in, zooming out, or aligning the entire screen where the diagram itself is located, is also provided. Completed experiments can be exported as an image, text file, or PDF document. The application uses two basic algorithms for calculations. The first type of calculation is done via the Louvain method based on modularity, and the second via ego zones. The work also explores the possibilities of expansion and future development of the application, which could include support for additional algorithms using virtual reality or support for mobile devices.

The result of this work is an innovative and user-friendly application that allows visualization and analysis of communities in social networks. As a result, social interactions and development in these types of networks can be examined. Using the newly created system, it was possible to monitor the behavior and development of communities and ego zones over time, which was further supplemented by detailed statistical information on the structure and dynamics of the entire network. With the help of several algorithms, various views of the network were obtained, which had to be displayed at different time intervals. The work also includes several freely available libraries, which together provide a comprehensive and clear system that is also applicable to larger networks.