Sentiment Analyzer Application Development

Portfolio project

Sentiment Analyzer

A modern sentiment analysis application that analyzes text from various sources including direct input, YouTube, e-commerce, news, social media, and blogs.

Overview

This application uses a custom sentiment analysis model to detect positive, negative, and neutral sentiments in text content. It features a React frontend and a Supabase Edge Function backend that performs the sentiment analysis using a custom implementation.

Features

Direct Text Analysis: Analyze any text input for sentiment

URL-based Analysis: Mock integration with various platforms:

YouTube videos

E-commerce product reviews

News articles

Social media posts

Blog content

Sentiment Visualization: Clear visual representation of sentiment with confidence scores

Responsive Design: Works on desktop and mobile devices

Technology Stack

Frontend:

React with TypeScript

TailwindCSS for styling

Vite for development and building

Backend:

Supabase Edge Functions (Deno runtime)

Custom sentiment analysis model

Sentiment Analysis Model

The application uses a custom sentiment analysis model that:

Identifies positive, negative, and neutral sentiments

Handles negations (e.g., "not good" as negative)

Processes intensifiers like exclamation marks

Provides confidence scores

Returns normalized sentiment scores from -1 (negative) to +1 (positive)

Installation

Prerequisites

Node.js 16+

npm or yarn

Deno (for local Edge Function development)

Supabase CLI (for deploying Edge Functions)

Setup

Clone the repository:

git clone https://github.com/yourusername/sentiment-analyzer.git
cd sentiment-analyzer

Install dependencies:

npm install

Set up environment variables:

cp .env.example .env.local

Edit .env.local with your Supabase URL and anon key

Run development server:

npm run dev

In a separate terminal, run the Edge Function locally:

cd supabase/functions/analyze-text
deno run --allow-read --allow-net --allow-env index.ts

Usage

Direct Text Analysis

Enter text in the "Text to Analyze" field

Click "Analyze Text"

View the sentiment results, including:

Sentiment classification (positive, negative, or neutral)

Confidence score

Visualization of sentiment score on a scale

URL Analysis (Mock)

Enter a URL in the input field

Select the appropriate category (YouTube, e-commerce, etc.)

View the aggregated sentiment analysis for the content

Browse through sample comments/reviews with individual sentiment ratings

Deployment

Frontend

Deploy to Vercel, Netlify or any static hosting service:

npm run build

Then deploy the contents of the dist folder.

Edge Function

Deploy to Supabase:

cd supabase/functions/analyze-text
supabase functions deploy analyze-text

Project Structure

/
├── src/                  # Frontend source code
│   ├── components/       # React components
│   ├── types/            # TypeScript type definitions
│   ├── utils/            # Utility functions and mock APIs
│   ├── App.tsx           # Main application component
│   └── main.tsx          # Entry point
├── supabase/
│   └── functions/
│       └── analyze-text/ # Edge Function for sentiment analysis
│           ├── index.ts  # Entry point for Edge Function
│           └── model.ts  # Sentiment analysis model implementation
├── public/               # Static assets
└── index.html            # HTML template

Future Improvements

Implement real API integrations for YouTube, news, and blog platforms

Add user authentication for saving analysis results

Enhance the sentiment model with more advanced NLP techniques

Add support for multiple languages

Implement aspect-based sentiment analysis

License

MIT License

Acknowledgements

Supabase for the serverless backend

React for the frontend framework