GROUP 1
• Rafael
• Eduardo Carvalho
• Livia Rabelo
Analyzing Factors Influencing Student Performance
A Data-Driven Approach to Educational Equity
Objectives
Provide actionable
recommendations to educators
and policymakers for fostering
equitable educational practices.
Getting to know the
Database
Source: Dataset from Kaggle (Fictional database for study
purposes)
Number of Samples: 6607
Attributes: 20 attributes (7 Numerical / 13 Textual)
● Academic Factors: Hours Studied, Attendance, Exam Score.
● Personal Factors: Sleep Hours, Motivation Level.
● Socioeconomic Factors: Family Income, Parental Education Level.
● Environmental Factors: Access to Resources, Distance from
Home.
● In the database, 235 empty cells were found, representing 4%
of the total database.
Preprocessing Steps
Data Cleaning: Imputation and standardization.
NULL Cells Attribute Quantity Attribute Quantity Attribute Quantity Total
Parental _Education_Level High School 42 College 18 Post Graduate 30 90
Distance_from_home Near 39 Moderate 21 Far 7 67
Teacher_quality High 23 Medium 46 Low 9 78
Adjustment Attribute Ocorrency % Attribute Ocorrency % Attribute Ocorrency % Total
Parental High 3223 49% College 1989 31% PostGraduate 1305 20% 6517
_Education_Leval School
Distance from home Near 3884 59% Moderate 1998 31% Far 658 10% 6540
Teacher_quality High 1947 30% Medium 3925 60% Low 657 10% 6529
Outlier Detection
Outliers: Managing anomalies to ensure robustness
Descriptive Statistics
The statistics about numeric variables.
Statistics Hours_Studied Attendance Sleep_Hours Previous_Scores Tutoring_Sessions Physical_Activity Exam_Score
MAX 44 100 10 100 8 6 101
AVERAGE 19.98 79.98 7.03 75.07 1.49 2.97 67.24
MIN 1 60 4 50 0 0 55
MEDIAN 20 80 7 75 1 3 67
MODA 20 67 7 66 1 3 68
STD DEV 5.99 11.55 1.47 14.40 1.23 1.03 3.89
RANGE 43 40 6 50 8 6 46
Exam Score Distribution
Female 67.24489796 Private 67.28770533
Male 67.22889355 Public 67.21291866
Parental Education Level
Parental Education Level
Parental Income
Parental Income
Parental Involvement
Parental Involvement
Hours Studied
Hours Studied
Hours of Sleep
Teacher Quality
Teacher Quality
Linear Regression
Questions
How do socioeconomic factors, such as family income and parental education
level, correlate with exam scores?
1. Family Income:
○ Positive association with exam scores.
○ Higher-income families' students outperform others.
2. Parental Education Level:
○ Positive association with exam scores.
○ Children of more educated parents outperform peers.
3. Performance Comparison:
○ Higher-income and more educated parents' children consistently outperform less
advantaged peers.
○ There are statistically significant differences.
Questions
How do personal habits, such as hours of sleep and hours studied, influence
academic performance?
1. Hours studied:
○ positively correlated with academic performance.
○ More study hours generally lead to better outcomes.
2. Sleep Hours:
○ Non-statistical relevant relationship with academic performance.
Questions
How significantly do school-related factors, such as teacher quality and school
type, affect exam scores?
1. Teacher Quality:
○ This is a strong and significant factor in academic performance.
○ Students with high teacher quality score consistently higher.
2. School Type:
○ No statistical significant influence on exam score.
Recommendations
For Policymakers:
● Invest in teacher training and mentorship.
● Provide performance-based incentives.
For Educators:
● Educate students about sleep and study schedules.
● Encourage physical activity for better focus.
For Communities:
● Offer free tutoring and scholarships.
● Engage parents through workshops.
Thank You