Student Performance Analysis and Prediction

Project Overview

This project analyzes a dataset of student exam scores to identify key factors influencing academic performance. The analysis involves data cleaning, in-depth exploratory data analysis (EDA), and the development of predictive models to forecast student scores.

A summary dashboard is also provided to visualize key findings and performance indicators.

Dataset

The dataset (StudentPerformanceFactors.csv) contains 6607 student records with the following attributes:

• Target Variable: Exam_Score
• Academic Factors: Hours_Studied, Attendance, Previous_Scores, Teacher_Quality,Tutoring_Sessions, Learning_Disabilities
• Home & Socio-Economic Factors: Parental_Involvement, Parental_Education_Level, Family_Income, Access_to_Resources, Internet_Access, Distance_from_Home
• Personal & Behavioral Factors: Extracurricular_Activities, Sleep_Hours, Motivation_Level, Peer_Influence, Physical_Activity, Gender
• School Factors: School_Type

Project Workflow

The project is structured in a Jupyter Notebook (Student Performance Factors.ipynb) and follows these key steps:

1. Data Cleaning and Preparation

• Checked for and handled missing values and duplicates.
• Standardized column names for consistency.

2. Exploratory Data Analysis (EDA)

Conducted in-depth EDA to understand the data distributions and relationships between variables:

• Univariate Analysis: Analyzed the distribution of individual features like gender, race/ethnicity, and parental education.
• Bivariate Analysis: Investigated the impact of various factors on exam scores, including:
  • School type vs. Scores
  • Distance from home vs. Scores
  • Gender vs. Scores
  • School type vs. Access to resources
• Correlation Analysis: A heatmap was generated to visualize the strong positive correlation between variables.

3. Predictive Modeling

The goal was to predict the Exam Score based on the available features.

• Preprocessing: Categorical features were converted into numerical format using One-Hot Encoding.
• Model Building: Two different regression models were built and evaluated:
  1. Linear Regression
  2. Lasso Regression
• Evaluation: Models were assessed using metrics such as R-squared (R2). A final comparison showed that all models performed similarly well, with high R2 scores.

4. Dashboard Visualization

A summary "Student Performance Analysis" dashboard (Dashboard.pdf) was designed to present the key findings in an accessible format.

• Key KPIs: Includes total students, gender distribution, and average/min/max scores.
• Key Insights: Features visualizations that show the relationship between scores and factors like gender, parental education, school type and test preparation.