Advanced Machine-Learning-Based Analysis of Critics and the Silent Majority Regarding the Energy Transition and Photovoltaics on TikTok

In this research project, several approaches of natural language processing are outlined to perform opining mining on the social media platform TikTok. The goal is to analyse the views of critics and the silent majority regarding the energy transition and photovoltaics. First, a dataset containing published TikTok videos with the hashtag #Energiewende, #Solarenergie and #Photovoltaik is created. After filtering out irrelevant content using an Ensemble Classifier model made up of a Random Forest and Support Vector Machine, the entities of the dataset undergo a sentiment analysis as well as a stance detection. Finally, the machine learning results are evaluated, revealing that there are slightly more videos in favour of the energy transition and photovoltaics. The engagement on the other hand is higher for neutral and negative videos. At last, the most used nouns by the critics are gathered and visualized with the aid of a word cloud. Although the word cloud does not show the context, in which the words are used, each word can be reviewed intuitively to gain a better idea of what aspects of the energy transition and photovoltaics are the most relevant to critics.