Student Performance Analysis Using
Power BI
Submitted by: [Your Name]
Course: [Course Name]
University: [University Name]
Date: [Submission Date]
1. Introduction
This report provides a comprehensive analysis of student performance using Power BI, a
powerful data visualization tool.
The aim is to understand how demographic variables and external influences such as
parental education and test preparation affect student scores across key subjects.
2. Problem Statement
Academic performance varies across individuals and groups.
Identifying trends and relationships between influencing factors and performance can help
stakeholders in making informed decisions to support students.
3. Dataset Description
The dataset includes the following columns:
- Gender
- Race/Ethnicity
- Parental Level of Education
- Lunch Type
- Test Preparation Course
- Scores in Math, Reading, and Writing
Source: Public student performance dataset from Kaggle.
4. Project Objectives
- Analyze score distribution by gender and parental education
- Examine the effectiveness of test preparation courses
- Compare performance across subject areas
- Develop a dashboard with interactive features for stakeholders
5. Tools & Technologies
- Microsoft Power BI Desktop
- DAX (Data Analysis Expressions)
- Data Modelling
- CSV file processing
- GitHub for version control
6. Data Cleaning & Transformation
Before importing the dataset into Power BI, it was cleaned using:
- Data type conversion
- Removal of missing values
- Standardization of categorical labels
7. Data Model
In Power BI, the cleaned CSV was imported as a single table.
Calculated columns and DAX measures were created for deeper insight.
No relationships with other tables were necessary for this analysis.
8. DAX Measures
Key DAX measures created:
- Overall Score = AVERAGE(math_score + reading_score + writing_score)
- Score Gap = math_score - reading_score
- Passed Test Prep = IF(test_preparation_course = "completed", 1, 0)
9. Dashboard Design
The dashboard was designed to include:
- Slicers for Gender, Education, Race
- Pie charts for completion rates
- Bar charts for subject comparison
- Card visual for average overall performance
10. Visual 1: Overall Scores by Gender
Shows average performance in math, reading, and writing grouped by gender.
A slight difference was observed in reading and writing scores.
11. Visual 2: Impact of Test Preparation
Compares scores of students who completed the course vs those who didn’t.
Completion correlates with improved performance in all subjects.
12. Visual 3: Influence of Parental Education
A bar chart showing average scores based on parents' highest education level.
Children of parents with higher education performed better on average.
13. Visual 4: Score Distribution by Lunch Type
Pie chart and stacked column chart showing students on free/reduced lunch scored lower
than those on standard lunch.
14. Visual 5: Racial/Ethnic Group Analysis
Displays how average scores differ across racial/ethnic groups.
Shows the need for equitable access and support for underperforming groups.
15. Insight Summary
- Students with completed prep courses scored significantly better
- Gender gaps were narrow, especially in reading
- Parental education strongly influenced performance
16. Limitations
- Dataset size limited to ~1000 students
- No socioeconomic or psychological factors considered
- No tracking of long-term academic performance
17. Recommendations
- Increase access to preparation resources for all students
- Focus support efforts on students with less parental educational support
- Conduct larger and longitudinal studies to validate findings
18. Future Work
Future improvements may include:
- Larger datasets
- Addition of behavioral metrics
- Integration with other educational dashboards
19. Conclusion
The Power BI dashboard successfully helped to identify and visualize key factors affecting
student performance.
With such tools, educators and policy-makers can better target their efforts.
20. References
- https://github.com/AhmedFathy-Power-BI-Analyst/Student-Performance
- Power BI Official Documentation
- Kaggle Student Performance Dataset
- DAX Guide (https://dax.guide)
6. Data Cleaning & Transformation (Expanded)
Effective data cleaning and transformation are critical steps in any data analysis project.
Before importing the dataset into Power BI, several preprocessing techniques were applied
to ensure the accuracy, consistency, and usability of the data. The quality of the final
dashboard and insights is directly dependent on the cleanliness and structure of the raw
data.
This section elaborates on the specific cleaning and transformation tasks undertaken in this
project, grouped into three major activities: data type conversion, handling missing values,
and standardization of categorical labels.
6.1 Data Type Conversion
When data is imported from a CSV file, Power BI automatically assigns data types to each
column. However, these automatic assignments are not always correct. In this project, each
column’s data type was manually reviewed and corrected if necessary to ensure proper
analytical behavior in Power BI.
For example:
- Numeric columns such as 'math_score', 'reading_score', and 'writing_score' were explicitly
set to the 'Whole Number' data type.
- Categorical fields such as 'gender', 'race/ethnicity', 'parental_level_of_education', and
'test_preparation_course' were converted to the 'Text' data type.
- Boolean fields such as 'test preparation completed' were encoded as binary values
(Yes/No or 1/0) to enable conditional formatting and filtering.
Proper data typing ensures that Power BI aggregates and visualizes the data accurately. It
also improves performance during model refresh and enhances compatibility with DAX
calculations.
6.2 Removal of Missing or Inconsistent Values
While the dataset was relatively clean, a thorough inspection was still conducted for any
missing, null, or inconsistent entries. In the real world, missing data can result from human
error, incomplete forms, or system limitations.
In this project, the following steps were taken:
- All rows with completely blank score entries were removed, as they do not contribute
useful information and may skew average values.
- Rows with partially missing data were examined for patterns. In cases where only one
score was missing, the row was retained if the other values were complete and useful.
- Inconsistent entries (e.g., numeric values in categorical fields) were corrected or excluded.
Ensuring data completeness and consistency is crucial, as even a small percentage of
corrupt records can impact the accuracy of DAX measures and visualizations.
6.3 Standardization of Categorical Labels
Categorical data is especially sensitive to inconsistencies in labeling. For instance, the same
category might be labeled in various ways, such as 'High School', 'high school', or
'Highschool'. These inconsistencies prevent accurate grouping and summarization in charts.
To address this:
- All categorical fields were reviewed for duplicates, case sensitivity issues, and spelling
errors.
- Labels were standardized to follow a consistent format. For example, all entries in
'parental level of education' were converted to Title Case (e.g., "Associate's Degree", "High
School", "Master's Degree").
- In 'race/ethnicity', group names like 'group A' were uniformly updated to 'Group A' for
presentation consistency.
- Binary responses such as 'yes' and 'no' in the 'test_preparation_course' column were
capitalized consistently.
This step greatly improved the interpretability and professionalism of the visualizations, as
legends and slicers reflected clean and readable labels.
6.4 Summary of Data Cleaning Process
The comprehensive data cleaning and transformation phase ensured that the dataset was
ready for advanced analysis and visualization in Power BI. Without this step, the reliability
of insights would be compromised, leading to misleading conclusions.
By performing data type conversion, removing incomplete records, and standardizing
category labels, the project established a strong foundation for accurate, meaningful, and
professional reporting.
