Project: “Social Media Sentiment Analysis” (Colab → UI + Deployment)
-
2 notebooks:
-
model development
-
playground/testing (.pkl, .h5)
-
-
Models:
-
Logistic Regression
-
Naive Bayes
-
ANN
-
-
Future:
-
multiple model versions
-
different optimizers/vectorizers/techniques
-
showcase all models in deployment
-
-
Stack:
-
Next.js
-
frontend/ + backend/
-
FastAPI backend
-
simple API flow
-
frontend fetch requests
-
-
Later additions:
-
Clerk auth
-
Supabase DB
-
user history storage
-
-
Goal:
-
scalable
-
clean architecture
-
conventional/best practices
-
-
Planned architecture:
- (inner files/folders may expand later)
social-sentiment/
├── frontend/ # Next.js App │ ├── app/ │ │ ├── page.tsx # Landing/Home │ │ ├── analyze/page.tsx # Main analysis UI │ │ ├── compare/page.tsx # Model comparison view │ │ ├── history/page.tsx # User history (after auth) │ │ └── api/ # Optional: lightweight proxy │ ├── components/ │ ├── lib/ │ └── package.json ├── backend/ # FastAPI │ ├── app/ │ │ ├── main.py # FastAPI app entry │ │ ├── models/ # Pydantic schemas │ │ ├── routers/ # API routes │ │ └── services/ # Business logic │ ├── ml/ # ML models & utils │ │ ├── models/ # .pkl, .h5 files (versioned) │ │ └── preprocessors/ # Vectorizers, etc. │ ├── requirements.txt │ └── Dockerfile # For containerization ├── docker-compose.yml # Local dev orchestration ├── .env.example # Environment template └── README.md
[NOTE: Lets go only one step at a time, don't give entire code. give even the terminal commands and working on external services.] Its Windows Device. What's done: Directory skeleton + git init. nothing else.