Critical_TSA: Twitter Sentiment Analysis in Critical Event Using Python

Twitter is a platform where users may post texts known as "tweets." They may be examined to determine their implications and meanings. Among these techniques, sentiment analysis is perhaps the most well-known. We addressed the issue of sentiment analysis of tweets on important occasions like disease or natural disasters. We called our proposal Critical_TSA, which means Twitter Sentiment Analysis (TSA) in critical event. We used SVM, to work on two datasets of distinct events at the psychological impact level: the first was tweets under the critical circumstances of the earthquake, and the second was the Covid-19 pandemic. Our results show that neutral tweets during the pandemic had the highest percentage in the analysis of the total tweets. We believe that the reason for this percentage being higher than others is that the impact was spread throughout the world. As for the earthquake, it is in one geographical area, so the number of positive earthquake tweets reached almost half of the total tweets. Finally, we implement the performance evaluation the SVM algorithm. Our proposed Critical_TSA reached a classification accuracy of up to 94% for tweets in the COVID-19 dataset and a classification accuracy of 92% for the earthquake dataset.