Microsoft 365 Sentiment Analyzer
Full-stack sentiment analysis across the Microsoft 365 ecosystem.
Microsoft 365 Sentiment Analyzer was designed and built by Yadnyesh Mulay, an AI-first full-stack developer based in Nashik, India. A modern full-stack web application (FastAPI + Next.js + TypeScript) that analyzes sentiment in communications across Teams, Outlook, OneDrive, SharePoint, and OneNote using a hybrid NLP engine (70% VADER + 30% TextBlob) with interactive Recharts visualizations and glass-morphism UI.
The problem
Organizations using Microsoft 365 lack accessible tools to analyze communication sentiment across Teams, Outlook, and other services without custom engineering.
The approach
Unified search interface across 6+ MS365 sources with hybrid NLP, interactive filtering, and executive-ready visualizations, deployable as a self-contained Docker stack.
How it was built
This project demonstrates full-stack engineering with a clean separation: FastAPI backend providing REST endpoints for authentication, search, analysis, and filtering; Next.js 14 frontend with App Router, TypeScript, Tailwind CSS, Framer Motion animations, and Recharts for interactive data visualization. The hybrid sentiment engine combines VADER (social media optimized) and TextBlob (general purpose) with weighted scoring.
The UI features a glass-morphism design system with smooth animations, responsive layouts, and accessibility compliance. Currently runs on dummy data generators simulating MS Graph API responses, real Microsoft Graph integration is spec'd for v1.1. Background task processing with live status polling provides real-time UX for long-running analyses.
Where the AI sits
Hybrid NLP ensemble (VADER + TextBlob), classical ML approach. Architecture designed for future LLM-based sentiment upgrade (BERT/GPT) per roadmap.
Stack
- Next.js 14
- TypeScript
- Tailwind CSS
- FastAPI
- Python
- VADER
- TextBlob
- Recharts
- Framer Motion
- Docker
Links
More work by Yadnyesh Mulay: all projects · about Yadnyesh Mulay · home