Student Performance Analysis

Project Overview

Key Questions

Which factors (gender, parental education, lunch, test prep) are associated with higher scores?
Is there correlation between math, reading, and writing scores?
Does completing test preparation improve outcomes?
Which student groups perform below expected levels and why?

Dataset

Primary dataset: Students Performance in Exams — Kaggle

Columns included: gender, race/ethnicity, parental level of education, lunch, test preparation course, math score, reading score, writing score.

Tools & Files

Tools

Excel — initial exploration & pivot tables
SQL (SQLite / PostgreSQL) — queries and aggregations
Power BI — interactive dashboard
Optional: Python (pandas) — for reproducible cleaning scripts

Data Cleaning Summary

Renamed all column names to Title Case.
Standardized Gender values — Female → F, Male → M.
Normalized text values in Race/Ethnicity, Parental Level of Education, Lunch, Test Preparation Course.
Handled missing values using Excel formulas and filters.
Added Average Score column with rounding.
Created Performance Category column to classify results.

Before Cleaning (Sample of 10 Rows)

gender	race/ethnicity	parental level of education	lunch	test preparation course	math score	reading score	writing score
female	group b	bachelor's degree	standard	none	72	72	74
male	group c	some college	free/reduced	completed	69	90	88
female	group b	master's degree	standard	none	90	95	93
male	group d	high school	standard	none	47	57	44
female	group c	associate's degree	free/reduced	none	76	78	75
female	group b	some college	standard	completed	88	95	92
male	group e	associate's degree	standard	none	59	66	63
female	group d	high school	free/reduced	none	40	52	43
male	group a	some high school	free/reduced	none	57	58	54
female	group c	some college	standard	completed	77	85	82

After Cleaning (Sample of 10 Rows)

Gender	Race/Ethnicity	Parental Level of Education	Lunch	Test Preparation Course	Math Score	Reading Score	Writing Score	Average Score	Performance Category
F	Group B	Bachelor's Degree	Standard	None	72	72	74	72.67	High
M	Group C	Some College	Free/Reduced	Completed	69	90	88	82.33	High
F	Group B	Master's Degree	Standard	None	90	95	93	92.67	High
M	Group D	High School	Standard	None	47	57	44	49.33	Low
F	Group C	Associate's Degree	Free/Reduced	None	76	78	75	76.33	Average
F	Group B	Some College	Standard	Completed	88	95	92	91.67	High
M	Group E	Associate's Degree	Standard	None	59	66	63	62.67	Average
F	Group D	High School	Free/Reduced	None	40	52	43	45.00	Low
M	Group A	Some High School	Free/Reduced	None	57	58	54	56.33	Low
F	Group C	Some College	Standard	Completed	77	85	82	81.33	High

Only first 10 rows are displayed for comparison before and after cleaning.

Data Analysis & Visualizations

Step 1: Student Performance by Gender

Comparison of average scores between Female (F) and Male (M) students using Power BI visuals.

Average Score by Gender Chart from Power BI

Step 2: Performance by Parental Level of Education

Clustered column chart and table visual in Power BI showing average scores by parental education level.

Average Score by Parental Education Level Chart in Power BI

Power BI Table Showing Average Scores by Parental Education Level

Step 3: Effect of Test Preparation Course

Comparison of students who completed the course vs. those who did not.

Average Score by Test Preparation Course

Number of Students by Test Preparation Course

Step 4: Performance by Race/Ethnicity

Average Score by Race/Ethnicity Chart in Power BI

Step 5: Performance by Lunch Type

Average Score by Lunch Type Chart in Power BI

Project Summary & Insights

Students whose parents have higher education levels tend to score higher.
Test preparation courses increase average scores by ~7 points.
Female students slightly outperform male students on average.
Students receiving standard lunch perform better than free/reduced lunch students.
Race/ethnicity shows performance gaps that may relate to external socioeconomic factors.