This thesis presents a recommender system for the Twitter and Instagram social networks.
The first chapters of the thesis are about the social networks and their marketing value, as well as about the system structure and the used technologies.
In chapters 3-4 the data collection, data processing and the difficulties what I have encountered during the development have been shown.
The stored data used to make statistical analysis which from I make a few general recommendations that can help improve the user interaction.
In Chapter 5 graphs are defined for more detailed data analysis, which are based on the relationships between different users.
The users have been clustered by a k-means algorithm and these clusters will be analyzed. Then the thesis focuses on how to find influencers, in the progress it is using a greedy set cover algorithm.
The thesis shows a way how to cut the hashtags and identify their themes. I have used link prediction in the section 5.5 to check which users will be related to each other in the future.
The thesis ends with a fast follower collector algorithm and a short presentation about the program general operation.