• Version
• Download 4
• File Size 4.96 MB
• File Count 1
• Create Date 2 October 2025
• Last Updated 2 October 2025

Enhancing User Experience and Undertaking Sentiment Analysis with Machine Learning in Social Media Twitter (X)

Submitted by

GK Sridharan

(Regd. No. 423206415017)

Under the guidance of

Prof. K. Venkata Rao

Head of the Department, CSSE

Department of Computer Science and Systems Engineering

Abstract
This project investigates the enhancement of user experience and engagement on the social media platform X (formerly Twitter) and Sentiment Analysis through the strategic application of advanced Machine Learning (ML) techniques. Despite its prominence as a communication tool for government agencies, policymakers, and influential figures—such as Heads of State—for disseminating critical announcements and shaping public perception during emergencies, Twitter (X) struggles with limited user engagement and lower adoption rates among the general public in many countries, including India, compared to platforms like Facebook and Instagram. This gap is primarily attributed to deficiencies in user experience, which this study seeks to address.

This study also undertakes Twitter Sentiment Analysis to demonstrate the application of machine learning in real-world social media data. Sentiment analysis helps classify tweets as positive, negative, or neutral, offering valuable insights into public mood, brand perception, and reactions to events. Using Python libraries such as Pandas, Scikit-learn, and TF–IDF vectorization, a supervised ML model was implemented and tested on the Sentiment140 dataset. The process involved data cleaning, feature extraction, and training classification models, which achieved reliable accuracy in distinguishing user opinions. This implementation showcases how machine learning can convert massive, unstructured tweet streams into actionable knowledge for businesses, researchers, and policymakers.

The research examines the current application of ML algorithms on X, focusing on features such as personalised content recommendations for a twitter user and undertaking sentiment analysis of a post with ML Model. Also, It identifies key challenges contributing to X’s suboptimal engagement and analyse the following challenges - Limited cross-platform integration (e.g., with WhatsApp) and possible solution and Text Length Restriction in a Tweet. Pros & Cons.

While policy-related concerns (e.g., phone-number-based authentication for new accounts) fall outside the study’s scope, the work emphasizes feasible, ML-driven solutions.

The findings and proposed models aim to bridge the gap between high-profile users (leaders, researchers, military organizations) and the general public, fostering a more interactive, inclusive, and user-centric ecosystem. By aligning Twitter’s design with evolving user expectations, this research positions X as a more dynamic and accessible social media platform.