🎓 Student Performance Predictor – Streamlit Deployment

A supervised machine learning web application built using Streamlit to predict students' final grades based on study habits, attendance, motivation, learning style, and other educational features.

---

🚀 Live Streamlit App

👉 Click below to use the deployed app:

🔗 https://laya123-star-student-performance-predictor-app-o3lju5.streamlit.app/

---

🚀 Run Notebook in Google Colab

Click below to open the notebook:

---

📘 Project Overview

This project builds and deploys a Machine Learning Classification Model to predict students' final grades.

The complete pipeline includes:

• Data Cleaning

• Exploratory Data Analysis (EDA)

• Feature Engineering

• Feature Scaling

• Feature Selection

• Model Building

• Hyperparameter Tuning

• Model Evaluation

• Model Deployment using Streamlit

Users can input student details and get real-time grade predictions.

---

🎯 Objective

The main objectives of this project are:

🔹 Analyze student performance data

🔹 Identify key factors affecting academic success

🔹 Build multiple ML classification models

🔹 Compare model performances

🔹 Optimize the best model using hyperparameter tuning

🔹 Deploy the model using Streamlit

---

📂 Dataset Description

Component	Description
Target Variable	FinalGrade
Problem Type	Multi-class Classification
Classes	A, B, C, D, F
Features	StudyHours, Attendance, Motivation, LearningStyle, Extracurricular, Resources, OnlineCourses, etc.

---

🧹 Data Preprocessing

The following preprocessing steps were performed:

✔ Loaded dataset using Pandas

✔ Handled missing values

✔ Removed duplicate records

✔ Detected and removed outliers

✔ Scaled numerical features using StandardScaler

✔ Split dataset into Training (80%) and Testing (20%)

---

📊 Exploratory Data Analysis (EDA)

Visualizations used:

• Histogram

• Box Plot

• Count Plot

• Heatmap Correlation

• Bar Plot

• KDE Plot

Insights:

• Understood feature distributions
• Identified outliers
• Analyzed feature relationships
• Checked class balance

---

🤖 Machine Learning Models Implemented

The following models were trained and evaluated:

• Decision Tree

• Random Forest

• k-Nearest Neighbors (k-NN)

• Support Vector Machine (SVM)

• Naive Bayes

• Logistic Regression

---

📊 Model Performance (Before Tuning)

Model	Accuracy	ROC-AUC
Decision Tree	0.878	0.919
Random Forest	0.824	0.946
k-NN	0.401	0.661
SVM (RBF)	0.352	0.606
Naive Bayes	0.287	0.532
Logistic Regression	0.284	0.526

✅ Best Model: Decision Tree

❌ Worst Model: Logistic Regression

---

⚙️ Hyperparameter Tuning

• Applied GridSearchCV

• Used Pipeline to avoid data leakage

• Tuned key parameters for better performance

---

📊 Model Performance (After Tuning)

Model	Accuracy	ROC-AUC	Best CV Score
Random Forest	0.901	0.986	0.859
k-NN	0.899	0.950	0.860
Decision Tree	0.878	0.919	0.839
SVM (RBF)	0.402	0.657	0.395
Naive Bayes	0.287	0.532	0.255
Logistic Regression	0.284	0.526	0.235

🏆 Best Model: Random Forest Classifier

• Accuracy: 90.1%

• ROC-AUC: 0.986

---

🏆 Final Model

✔ Random Forest Classifier selected

✔ Optimized using hyperparameter tuning

✔ Saved using Pickle for deployment

import pickle

pickle.dump(best_rf_model, open("Models/best_rf_model.pkl", "wb"))

---

📊 Evaluation Metrics

• Accuracy Score

• Confusion Matrix

• Precision

• Recall

• F1-Score

• ROC-AUC Score

• Feature Importance

---

🌐 Streamlit Deployment

The trained model is deployed using Streamlit.

Features:

✔ User-friendly interface

✔ Real-time predictions

✔ Input student data easily

✔ Instant grade prediction output

---

⚠️ Limitations

• The model depends on the quality and completeness of the dataset

• Limited features may not capture all real-world academic influence

• Performance may vary for students from different educational systems

• Does not include psychological or external environmental factors

• Model may not generalize well to unseen or highly imbalanced data

---

🛠 Tech Stack

Tool	Purpose
Python	Programming language
Pandas	Data handling
NumPy	Numerical computation
Scikit-learn	ML models & preprocessing
Matplotlib / Seaborn	Visualization
Pickle	Model saving
Streamlit	Web app deployment
Google Colab	Development

---

📁 Repository Structure

student-performance-predictor/

│

├── app.py

├── Models/

│ └── best_rf_model.pkl

├── notebook.ipynb

├── README.md

---

🚀 How to Run the Project

1️⃣ Clone Repository

git clone <your-repo-link>
cd student-performance-predictor

2️⃣ Install Dependencies

pip install -r requirements.txt

3️⃣ Run Streamlit App

streamlit run app.py

---

📌 Academic Submission

This repository was developed as part of a Machine Learning & Data Science program to demonstrate data preprocessing, exploratory data analysis (EDA), model building, hyperparameter tuning, and deployment of a machine learning model using Streamlit for real-time student performance prediction.

---

👤 Author

Name: Laya Mary Joy

Organization: Entri Elevate

Date: February 14, 2026

---

⭐ Acknowledgment

Thanks to Entri Elevate for guidance and support in building this project.

---

📌 Future Improvements

• Add more student behavioral features

• Improve model generalization

• Deploy using Docker / Cloud platforms

• Add visualization dashboard

---