Student ML Project

A simple Machine Learning mini-project that predicts a student's performance level (Low, Medium, High) based on Age and Marks.

This repository contains two parts:

student-ml-project │ ├── api → Flask-based Machine Learning API └── ui → React-based Frontend UI

This is one of my early ML learning projects, built using a very small sample dataset (5 records) just to understand the end-to-end workflow of ML + API + UI + Hosting.

LIVE DEMO

Backend API (Render): https://student-ml-api.onrender.com/predict?age=35&marks=89

Frontend UI (Render): https://student-ml-ui.onrender.com/

(Note: Render free tier sleeps when idle.)

PROJECT STRUCTURE

student-ml-project/ │ ├── api/ │ ├── main.py │ ├── requirements.txt │ ├── label_map.pkl │ ├── normalization_values.pkl │ └── student_data_cleaned_for_ML.pkl │ └── ui/ ├── public/ ├── src/ ├── package.json └── .gitignore

MACHINE LEARNING DETAILS

• Model trained using only 5 sample records (for learning purposes) • Labels: 0 → Low 1 → Medium 2 → High

• Preprocessing: Min–max normalization • Model: Basic classifier (demonstration level)

This model is NOT accurate — it is only for learning how to train, save, load, and integrate ML with API & UI.

TECHNOLOGIES USED

Backend (API): • Python 3 • Flask • Pickle • Render (Hosting)

Frontend (UI): • React (Create React App) • HTML, CSS, JavaScript • Fetch API • Render (Deployment)

UI FEATURES

• Enter Age and Marks • Sends request to ML API • Displays: – Predicted label – Probabilities (Low, Medium, High)

(No input validations added — kept simple for learning.)

API ENDPOINT

GET /predict?age=VALUE&marks=VALUE

Example: https://student-ml-api.onrender.com/predict?age=35&marks=89

Sample Response: { "age": 35.0, "marks": 89.0, "predicted_class": 2, "predicted_label": "High", "probabilities": { "Low": 0.0, "Medium": 0.23, "High": 0.77 } }

HOW TO RUN LOCALLY

Running the API

Step 1: cd api

Step 2: pip install -r requirements.txt

Step 3: python main.py

API available at: http://127.0.0.1:5000/predict?age=20&marks=80

Running the React UI

Step 1: cd ui

Step 2: npm install

Step 3: npm start

UI available at: http://localhost:3000/

FUTURE IMPROVEMENTS

• Train with larger dataset • Add validations in UI • Improve model accuracy • Better error handling • Charts / confidence visualization in UI • Dockerize the project

PROJECT PURPOSE

This is a learning project created to understand:

• Training and saving ML models • Building APIs using Flask • Connecting frontend to backend • Deploying APIs and UIs • Organizing code in GitHub