The number of user-generated text data grows exponentially as social networks, blogs, forums, and commercial websites become more widely used. For manufacturers, retailers, and suppliers of products and services, user feedback on a product or service review is very important. As a result, sentiment analysis and opinion mining for decision-making purposes play a very crucial role. In this work, two datasets from different domains are used. The sentiment of each review in all the datasets is identified using the VADER sentiment analyzer. Results from feature-based sentiment analysis represent a significant advancement in product design, product analysis, and product market share. Topic-based modeling methods are important for extracting product features from a large amount of raw data, including consumer reviews, social media comments, etc. The aspect category in this work is extracted using Latent Dirichlet Allocation (LDA). The model is built to provide leads from user reviews for product design and quality assurance. The summary of each extracted topic is generated, and its performance is evaluated in terms of accuracy. The accuracy achieved by the work is 60%.