Sentinic
Project
Sentinic
Date
Oct 2021
Overview
Sentinic is a dynamic web-based application designed to scrape online reviews and harness the power of machine learning models for precise sentiment analysis.
01 Features included
Authentication Options: Seamlessly log in with Google credentials or choose to use the platform anonymously for a hassle-free experience.
Product Selection: Explore a wide range of products using the dropdown menu, ensuring diverse options for sentiment analysis.
Custom Product Analysis: Alternatively, input Amazon product URLs of your choice for in-depth review scraping and sentiment analysis.
Review Scraping: Harness the power of Sentinic to scrape detailed Amazon product reviews for insightful analysis.
Machine Learning Insights: Benefit from accurate sentiment predictions derived from machine learning models, providing a deeper understanding of product sentiment.
User Review Access: Access and explore user reviews related to your selected products, gaining valuable insights into customer feedback.
Monthly Sentiment Trends: View counts and sentiment percentages across months for your chosen products, helping track sentiment fluctuations over time.
Global Brand Growth: Get a comprehensive overview of your brand's growth through global sentiment percentage analysis.
View website02 development
* In the development of 'Sentinic,' a versatile web-based application, I served as a full-stack developer, creating a dynamic and user-friendly interface using React, Redux, and MUI for the front-end. On the back-end, I expertly leveraged Node.js, Express, and MongoDB, establishing a robust foundation for data management. To optimize performance and enable efficient job sequencing, I implemented Bull along with Socket.io, facilitating seamless data exchange without user interaction. For sentiment analysis, 'Sentinic' employs three powerful algorithms—Logistic Regression, Support Vector Machine, and Random Forest—to ensure precise sentiment classification. The models are efficiently deconstructed using 'pickle,' enabling them to seamlessly run within the Node environment. This comprehensive development approach results in a reliable and efficient platform for online review analysis.
View code