LaptiQ

• 

  Front Page

Inspiration

Buying a laptop is overwhelming with so many options, specs, and mixed reviews. Users often get confused about whether a laptop’s performance or battery life suits their needs, and they struggle to estimate if the price is fair. We wanted to build an AI-powered assistant that not only analyses reviews intelligently but also predicts laptop prices, making purchase decisions easier and more data-driven.

What it does

LaptiQ is a web application that:

Allows users to submit laptop reviews with ratings

Uses NLP-based sentiment analysis to detect if a review is positive or negative

Extracts aspect-based sentiments (e.g. battery, display, performance) to show what users liked or disliked

Predicts the approximate price of a laptop based on specifications using a trained ML regression model

Provides a dashboard summarising reviews, recent sentiments, and product ratings for quick insights.

How I built it

Frontend: HTML, CSS, JavaScript (jQuery) for dynamic forms and AJAX requests

Backend: Flask framework with user authentication using Flask-Login and SQLAlchemy ORM

Machine Learning Models:

Sentiment analysis: Trained Logistic Regression model on review text with TF-IDF vectorization

Price prediction: Regression pipeline using Random Forest trained on laptop specs dataset

Aspect extraction: Implemented with TextBlob polarity analysis on detected aspect keywords

Deployment: Hosted on Render with persistent CSV data storage for reviews

Challenges Iran into

NLTK punkt download errors during deployment and container builds

Extracting meaningful aspects from short reviews without false positives

Handling inconsistent product URLs from Amazon for automatic company and model extraction

Ensuring models trained locally work seamlessly on server environments with different scikit-learn versions

Accomplishments that I am proud of

Built a full-stack ML-powered assistant integrating NLP and regression models

Successfully deployed a working Flask app with user authentication and real-time model inference

Implemented aspect-based sentiment analysis that provides deeper insights beyond generic review polarity

Designed a clean, intuitive UI for users to submit reviews and predict laptop prices effortlessly

What I learned

Integrating multiple ML models into a single production-ready web app

Handling version compatibility issues when deploying scikit-learn models

The importance of aspect-level analysis in reviews to derive actionable insights

Practical skills in full-stack deployment, user session management, and pipeline structuring

What's next for LaptiQ

Upgrade to LLM-based aspect extraction for richer and more nuanced insights

Add a recommendation engine suggesting laptops based on user preferences and past reviews

Implement a comparison feature between laptops based on specs, price, and aggregated sentiments

Develop an admin dashboard for brands to analyse customer feedback trends

Build a mobile-friendly UI and deploy on cloud with scalable database integration