Analyzing and predicting trends in contemporary social discourse through hashtag campaigns

Abstract

In the current landscape of social media, hashtags play a significant observable role in boosting the movement of information across a diverse range of fields and people. This research aims to identify and then examine extremely popular hashtags on social media and uncover the characteristics that make these hashtags popular. It attempts to follow the hashtag along its journey over time in gaining and losing popularity and in doing so, extrapolates to analyze the boom of past trends over time and the impact that hashtags have in propagating information. The research further tries to create a structure for a network showcasing the propagation of information and interaction between users as they engage using the hashtag in an attempt to understand why and how these hashtags reach such a diversified range of groups and people. It also attempts to be able to predict and forecast the trend and direction of the hashtag and the interaction it generates after the interaction period. Using a mixture of modern techniques such as network science and graph theory concepts, unsupervised machine learning tools, greedy modularity maximization model, and many other natural language processing (NLP) tools such as LSTM, there is the potential to understand the influence of popularity through hashtags and be able to predict the popularity of hashtags.