The most talked about topic of interest in the medical realm as of today, is the debate on the impact that COVID-19 vaccine has on individuals, and their response in encountering the virus. While there are quite a few vaccine variants that have been developed, there has always been a lingering ambiguity in declaring that an individual can be completely immune to the virus. There have been many studies whilom this cognition of analysing the sentiment perception of vaccines, however the data utilization from various sources and the apropos implementation using the language processing methodologies have lagged a great deal. This paper pivots on the data drawn from social media platforms, and optimizes the sentiments using the Natural Language processing Toolkit (NLTK). The process of word embedding, with TFIDF vectorizer commingled with data unsheathing through fine-grained sentiment analysis and machine learning algorithms such as Linear SVC, SVM and Naïve bayes on the covid19 dataset have aided in stratifying the public tweet sentiments based on their polarity, precision, recall, f1-score value and support. The simulations have been implemented using the lexicon, rubric-based analytical tool VADER (Valence Aware Dictionary and sentiment Reasoner) incorporated in Python specifically for optimized extraction of sentiments from data.