Unveiling Hinglish Sentiment Trends on MakeMyTrip: VADER vs SATV Analysis

Information is extracted from data using statistical and analytical techniques in the fields of data science. This broad area encompasses a variety of operations, including business intelligence, data visualization, sentiment analysis, and statistical modeling. Data scientists use these techniques to identify trends and patterns in data, forecast results, and support decision-making. This study analyzes the opinions of MakeMyTrip users by extracting data from Twitter X using Python-based scraping scripts. A dataset of approximately 5,000 tweets was collected for sentiment analysis, categorizing customer sentiments into positive, negative, and neutral. Three machine learning classifiers Support Vector Classifier (SVC), Logistic Regression, and Naïve Bayes were employed to assess sentiment polarity and accuracy. The proposed Sentiment Analysis Translating Vader (SATV) model was implemented to determine sentiment polarity and accuracy, followed by the application of a confusion matrix to compute Precision, Recall, and F-measure. Additionally, sentiment polarity was analyzed using VADER, facilitating a comparative evaluation between the SATV model and VADER. The results indicate that the SATV model outperforms VADER in terms of accuracy, precision, recall, and F-measure. Experimental findings reveal that negative tweets frequently contain terms such as “trip,” “hotel,” “difficult,” “time,” and “food.” Addressing these concerns could enhance user trust and satisfaction with MakeMyTrip. Furthermore, this approach can contribute to service improvements and aid in predicting customer satisfaction, thereby supporting the platform’s growth and service optimization.