Analysing the Social-Economic Impact of Wireless Mobile Services During and Before COVID-19 Using Topic Modelling and Sentiment Analysis on Tweets

Abstract

Social media platforms can be used as a data source for measuring public opinion on various topics such as wireless mobile services. Twitter is a suitable platform that is able to map the sentiments. In this research the influence of wireless mobile services on values such as user satisfaction (social effect), affordability (economic effect) and willingness (social effect) is researched. This research is conducted through a created system that uses topic modeling and sentiment analysis. HDP, LDA and LSI are the topic models used to map the various topics. While Multinomial Logistic Regression, Naive Bayes, Decision Trees and Random Forest are the sentiment models that map the sentiment per value. All these models are evaluated for their performance with the aim of choosing the best model for the system. This research will also determine the sentiment over time for each value and the sentiment for the companies Mint Mobile and Infinity Mobile. These companies have different policies, the aim of this analysis will therefore be to provide insight into the effect of company policy on user satisfaction, affordability and willingness. The analysis has shown that the overall sentiment for user satisfaction, affordability and willingness is negative. This research also showed that the pandemic has played a major role in this negative sentiment. For both willingness and affordability, a clear trend break can be observed at the start of the pandemic. Finally, it is also observable that for all values , the company with a more flexible policy(Mint Mobile) has a less negative sentiment than the company (Infinity Mobile) with a traditional policy.