Designing an information system for integrated topic analysis of social media big data

Open communities of users in social media are a source of data that quickly presents the thematic agenda of issues relevant to the population. The indicators of user activity are views, likes, comments and reposts, and they are of a dynamic nature. The article presents a new vision at the topic modeling problems, the results of which are examined for dynamic properties. These data are relevant to solve problems of information support for regional and municipal development. The authors reveal their experience in designing an information system for integrated topic analysis of large open social media data. The system is based on three technologies: building dynamic topic models for monitoring social media, intelligent analysis of topic modeling results; and cognitive visualization of dynamic topic modeling results. To take into account design uncertainty, object modeling tools, system design and a modular approach were used.