In order to correctly identify categories we analyze bio and content of influencers' posts. We use methods of mathematical statistics and machine learning to understand their meaning and, based on this, indicate the appropriate topics. At the moment we have more than 130 topics in our database.
The results are checked daily by a team of trained QA and are used to further train the topic detector. In addition, influencers themselves can suggest edits to the topics on the reports, and this data is viewed by our QA and used for the additional training of the algorithm.
To learn how to change the topic please check out this article.