Evaluation of Indian EVs based on FI attributes using Al/ML in Collaboration with Mahindra Research Valley

This project is aimed at analyzing the performance of different types of vehicles based on various parameters such as Range, Ride, Handling, Braking, Steering, Off-road capability, NVH, Occupant Comfort, Visibility & Package, Safety, and Software Features. We have used web scraping to collect reviews from various car review sites and used sentiment analysis to determine the score for each parameter based on the keywords used in the reviews.

  • Developed a vehicle performance analysis project to compare different types of vehicles based on various parameters such as Performance, Range, Ride, Handling, Braking, Steering, Off-road capability, NVH, Occupant Comfort, Visibility & Package, Safety, and Software Features.
  • Implemented web scraping techniques to gather car reviews from multiple sources, including car review sites like Cardekho, Zigwheels, and Carwale.
  • Utilized the VADER (Valence Aware Dictionary and sEntiment Reasoner) algorithm from the NLTK library for sentiment analysis of the collected reviews.
  • Analyzed the sentiment of the reviews to determine the score for each parameter using keyword-based analysis.
  • Normalized the scores for each parameter to a range of [0, 1] for better comparison and aggregation.
  • Created an Excel output with the metric scores for each car, providing a comprehensive performance comparison.
  • Demonstrated strong data processing and analysis skills, handling large amounts of text data and extracting valuable insights.