Social media has emerged as the most important platform for individuals, organizations, and governments to communicate their views in the world. The same proves to be crucial that social media sites have played during the pandemic of coronavirus disease 2019 (COVID-19), wherein people communicated, shared, and expressed what they perceived. Additionally, betterment of response to such loads pertaining to time-critical issues can be done through analysis of such textual data. Text data preprocessing includes cleaning, tokenizing the text, and stemming followed by the conversion of text data into numerical vectors through CountVectorizer. The Logistic Regression and SVC algorithms were used. The data is then fitted into the SGDClassifier model and further prepared for evaluation against accuracy scores and classification reports. The best result was achieved with the SGDClassifier algorithm with CountVectorizer, with an accuracy of 88%, a precision of 87.9%, a recall of 87.9%, an F1-score of 87.8%, and AUC of 86%, respectively. The results of the study highlight the significance of class-imbalance management and prove that machine learning techniques are potent to decide the public sentiment at the real-time level. This work will serve as a benchmark for the future in the promising area, and it is in line with how the social media data can be utilized to take into account public sentiments in a pandemic situation.