COVID-19’s Impact on the Telecommunications Companies

Abstract

Now the world is witnessing most significant challenges due the Covid-19 crisis. Beyond health effects, it has social and economic effects. With the enormous amount of data available and the widespread use of social web globally, research can and should use it to provide solutions. Customer satisfaction is known to affect customer churn (customers leaving companies), which is a problem affecting many companies in competitive and volatile markets – like the current one. One easily available open source of customer opinions are tweets – more relevant now in the online world. Whilst Natural Language Processing (NLP) on tweets is not new, few studies target customer satisfaction, and NLP body of research on Arabic tweets is modest; we are not aware of any other study on this during a global pandemic. Our research thus aims to propose a new model based on Twitter mining to measure customer satisfaction during Covid-19, as well as compare customer satisfaction before and during the crisis. This is a use case for the largest Telecom companies in Saudi Arabia, and we involve the popular method of Sentiment Analysis (SA) for the task. We additionally propose a new Saudi lexicon and apply it to monitor real-time customer satisfaction on Twitter using three different transfer network models on Arabic sentiment analysis. Also, this research evaluates using these models on Arabic Sentiment Analysis as the first study comparing between three different transfer network models on Arabic text.