Categories: Power BI
Tags:

Student Performance Analysis in Excel: Step-by-Step Guide to Build an Interactive EDA Dashboard

Introduction

When people talk about analytics dashboards, they usually jump straight to Power BI or Python. But over the years, while training thousands of students and professionals, I’ve seen one truth repeat itself again and again — Excel is still one of the most powerful tools for exploratory data analysis (EDA) when used correctly.

In this tutorial, I’ll walk you through how I created the Student Performance Classification – EDA Dashboard entirely in Microsoft Excel. This dashboard is not just about charts; it’s about understanding patterns behind student behavior and translating raw data into meaningful insights that educators, institutions, and analysts can actually use.

Get the dataset here: https://github.com/slidescope/Student-Performance-Classification-Dataset-for-DA-and-ML-Practice-by-Slidescope

The objective of this analysis is simple:
👉 Understand what factors influence student performance levels (High, Medium, Low).

Using variables like study hours, attendance percentage, assignments completed, practice score, sleep hours, and screen time, we explore how lifestyle and academic habits directly affect outcomes. Instead of jumping into machine learning straight away, this dashboard focuses on EDA — the most critical step before any predictive modeling.

What I like about this use case is its realism. Every column reflects a real-world variable that teachers, parents, and students care about. And by building this in Excel, we make the analysis accessible to beginners while still applying analytical thinking used by professional data analysts.

In this guide, I’ll explain:

  • How to understand and structure the dataset
  • How performance levels are analyzed using averages and variance
  • How different Excel visuals reveal hidden patterns
  • Why this kind of dashboard is a strong foundation for future ML models

If you’re learning data analytics, Excel dashboards like this will sharpen your business thinking, not just your technical skills. Let’s break it down step by step.

Dataset Explanation

Before building any dashboard, the most important step is understanding the dataset. No visualization or KPI makes sense unless you know what each column represents and how it connects to the problem you’re solving.

This dataset focuses on student performance analysis, where the final outcome is a categorical variable called performance_level with three classes:

  • High
  • Medium
  • Low

Let’s go through each column and understand why it matters.

1. student_id

This is a unique identifier for each student.

  • Purpose: Identification only
  • Usage in analysis:
    Not used in calculations or charts, but essential for maintaining row-level uniqueness.

In dashboards, IDs are usually excluded from visuals but remain important if you later export filtered data or connect to another dataset.

2. study_hours

Represents the average number of hours a student studies per day.

  • Type: Numeric (continuous)
  • Why it matters:
    Study time is one of the strongest predictors of academic performance.

In the dashboard:

  • You can clearly see that High-performing students have higher average study hours
  • Low performers consistently show fewer study hours

This metric is ideal for:

  • Bar charts by performance level
  • Correlation analysis with practice score

3. attendance_pct

This shows the percentage of classes attended by the student.

  • Type: Numeric (percentage)
  • Why it matters:
    Attendance reflects discipline, consistency, and exposure to learning.

In the dashboard:

  • High performers show significantly higher attendance percentages
  • Low performers cluster around lower attendance values

This variable is crucial because it often acts as a behavioral indicator, not just an academic one.

4. assignments_completed

Represents the number of assignments completed by the student.

  • Type: Numeric (discrete)
  • Why it matters:
    Assignment completion reflects engagement and responsibility.

In the combined bar + line chart:

  • High performers complete more assignments
  • Practice score rises alongside assignment completion

This variable works extremely well in dual-axis charts, helping compare effort vs outcome.

5. practice_score

This is the average score obtained in practice tests or assessments.

  • Type: Numeric
  • Why it matters:
    This is a direct proxy for academic preparedness.

From an analytical perspective:

  • Practice score strongly correlates with performance level
  • It also connects indirectly with study hours and attendance

This column helps validate whether effort (study, assignments) is translating into results.

6. sleep_hours

Represents the average number of hours a student sleeps per day.

  • Type: Numeric
  • Why it matters:
    Sleep is often ignored in academic analysis, but it plays a critical role in cognitive performance.

In the dashboard:

  • Extremely low or extremely high sleep hours are associated with lower performance
  • Medium-to-high performers show more balanced sleep patterns

This insight is powerful because it shows non-academic factors influencing outcomes.

7. screen_time

This indicates daily screen usage hours, including mobile, TV, or computer time unrelated to study.

  • Type: Numeric
  • Why it matters:
    Excessive screen time often correlates with distraction and reduced focus.

In the dashboard:

  • Variance in screen time is highest for low-performing students
  • High performers show more consistent (lower variance) screen usage

Instead of just averages, variance analysis here helps identify behavioral instability.

8. performance_level (Target Variable)

This is the classification label assigned to each student.

  • Categories: High, Medium, Low
  • Why it matters:
    This is the outcome we are trying to understand and eventually predict.

In the dashboard:

  • All KPIs and visuals are segmented by performance level
  • Slicers allow instant comparison across groups

This variable acts as the lens through which all other features are evaluated.

Why This Dataset Is Ideal for EDA

This dataset is extremely well-suited for exploratory analysis because:

  • It has both academic and lifestyle variables
  • The target variable is already classified
  • Relationships are intuitive but still need validation
  • It prepares the ground for classification models later

From an Excel dashboard perspective, it allows:

  • KPI cards for averages
  • Bar charts for categorical comparisons
  • Scatter plots for relationship analysis
  • Variance analysis for behavioral insights

From EDA to Machine Learning (Conceptual Bridge)

After completing this dashboard, the next logical step would be:

  • Encode performance_level
  • Normalize numeric features
  • Use models like Logistic Regression, Random Forest, or XGBoost

But without this EDA step, any ML model would be blind.

This Excel dashboard answers the “why” before we move to the “how.”

That’s the mindset I always recommend:
👉 Understand the story first. Predict later.

How I Built the Student Performance Classification EDA Dashboard in Excel

A Step-by-Step Practical Tutorial

When I design dashboards, I don’t start with charts — I start with questions.

For this student dataset, my core questions were:

  • What differentiates High, Medium, and Low performers?
  • Which habits actually matter?
  • Are lifestyle factors (sleep, screen time) as important as academic ones?

Once these questions are clear, the dashboard structure almost designs itself.

Let me now walk you through exactly how each section of this Excel dashboard was built, step by step.

STEP 1: Preparing the Dataset (Foundation Layer)

1. Importing the Data

  • The dataset was loaded into Excel as a structured table
  • Shortcut: Ctrl + T
  • Table name used: StudentData

Why this matters:

  • Structured tables auto-expand
  • Formulas become readable
  • PivotTables update instantly

2. Data Validation & Cleanup

Before any visualization:

  • Checked for blanks
  • Ensured numeric columns were numbers
  • Ensured performance_level had only:
    • High
    • Medium
    • Low

No transformations were forced at this stage — this is pure EDA, not modeling.

STEP 2: KPI Cards (Top Summary Metrics)

KPIs Created:

  • Avg study_hours
  • Avg attendance %
  • Avg assignments_completed
  • Avg practice_score
  • Avg sleep_hours
  • Avg screen_time

How Each KPI Was Built

  1. Insert → PivotTable
  2. Rows: empty
  3. Values: target metric (Average)
  4. Filter: performance_level (connected later to slicer)

Example:

Average of study_hours

Each KPI PivotTable was:

  • Converted into a single-value cell
  • Styled using:
    • Shapes
    • Icons
    • Bold typography

This gives the executive summary view of student behavior.

STEP 3: Performance by Avg Study Hours (Horizontal Bar Chart)

Question Answered:

Do high performers actually study more?

Steps:

  1. PivotTable
    • Rows: performance_level
    • Values: Average of study_hours
  2. Insert → Bar Chart (Horizontal)
  3. Sorting:
    • Descending by average study hours

Why Horizontal?

  • Performance levels are categories
  • Easier visual comparison

Insight:

High performers clearly spend significantly more study hours, validating a foundational assumption.

STEP 4: Performance by Attendance % (Column Chart)

Question:

Is attendance a strong performance indicator?

Steps:

  1. PivotTable
    • Rows: performance_level
    • Values: Average of attendance_pct
  2. Insert → 3D Column Chart (used for visual emphasis)
  3. Formatting:
    • Data labels ON
    • Clean axis titles

Insight:

Attendance drops sharply for Low performers.
This tells us discipline matters as much as intelligence.

STEP 5: Assignments Completed vs Practice Score (Combo Chart)

Question:

Does effort translate into results?

Steps:

  1. PivotTable
    • Rows: performance_level
    • Values:
      • Average of assignments_completed
      • Average of practice_score
  2. Insert → Combo Chart
    • Assignments → Column
    • Practice Score → Line (Secondary axis)

Why Combo Chart?

  • Shows effort vs outcome together
  • Perfect for storytelling

Insight:

Low performers show both:

  • Fewer assignments
  • Lower practice scores

This confirms a behavior-performance loop.

STEP 6: Performance by Sleep Hours (Pie Chart)

Question:

Does sleep impact performance?

Steps:

  1. PivotTable
    • Rows: performance_level
    • Values: Average of sleep_hours
  2. Insert → Pie Chart
  3. Color coding:
    • High: Calm color
    • Medium: Neutral
    • Low: Alert color

Insight:

Balanced sleep appears in higher performance groups.
Too much or too little sleep correlates with lower outcomes.

This is a non-obvious but powerful insight.

STEP 7: Variance of Screen Time (Behavior Stability Analysis)

Question:

Is screen time consistency important?

Instead of average, I analyzed variance.

Steps:

  1. Helper PivotTable
    • Rows: performance_level
    • Values: VAR.P of screen_time
  2. Insert → Horizontal Bar Chart

Why Variance?

  • Average screen time can be misleading
  • Variance shows behavior stability

Insight:

Low performers show highest screen-time volatility, indicating distraction and lack of routine.

This is advanced EDA thinking, not basic Excel.

STEP 8: Attendance % vs Practice Score (Scatter Plot)

Question:

Are attendance and practice score correlated?

Steps:

  1. Insert → Scatter Plot
  2. X-axis: attendance_pct
  3. Y-axis: practice_score
  4. Color-coded by performance_level

Why Scatter?

  • Shows relationship, not totals
  • Highlights clusters

Insight:

High performers cluster in the top-right quadrant
Low performers scatter widely — inconsistency again.

This chart visually proves correlation without statistics.

STEP 9: Performance Level Slicer (Interactivity Layer)

Steps:

  1. Insert → Slicer
  2. Field: performance_level
  3. Connected to:
    • All PivotTables
    • All KPI cards

Why Slicers Matter:

  • One-click exploration
  • Makes Excel feel like Power BI
  • Essential for decision-makers

STEP 10: Dashboard Layout & Design

Canvas Background

  • Designed in Canva
  • Exported as PNG
  • Inserted as background image

Layout Principles Used:

  • KPIs on top
  • Behavior charts in middle
  • Relationship charts at bottom
  • Left-to-right storytelling

Why Design Matters:

Good analysis with bad design = ignored insights.

STEP 11: Dashboard Thinking (Most Important Part)

This dashboard answers:

  • What matters most?
    Attendance, study hours, consistency
  • What matters less than assumed?
    Raw screen time averages
  • What’s actionable?
    Improve habits, not just content

This is why EDA is not about charts, it’s about decisions.

Final Thoughts from Ankit

If you can build this dashboard in Excel, you are already thinking like:

  • A data analyst
  • A business decision-maker
  • A future ML practitioner

Before models, before Python, before AI —
EDA is where real understanding happens.

Conclusion

This Student Performance Classification EDA Dashboard demonstrates how powerful Excel can be when it is used with the right analytical mindset. Many people underestimate Excel, but as this dashboard proves, it is more than capable of delivering deep insights, interactivity, and professional-level analysis when structured correctly. The key is not the tool itself, but how you think while using it.

Through this analysis, we clearly saw that student performance is not driven by a single factor. High-performing students consistently show stronger habits—better attendance, higher study hours, more completed assignments, and stable daily routines. On the other hand, low performance is strongly linked with inconsistency, irregular screen-time behavior, lower engagement, and weaker practice outcomes. These insights are far more valuable than raw averages because they help educators and institutions focus on behavioral improvements rather than surface-level metrics.

This dashboard also highlights the importance of exploratory data analysis before moving toward predictive modeling. By visually understanding relationships, distributions, and patterns, we reduce guesswork and make smarter decisions when selecting features for machine learning models. In fact, fields like attendance percentage, study hours, practice score, and screen-time variance clearly emerge as strong predictors of performance level.

From a learning perspective, building this dashboard strengthens multiple skills at once—data cleaning, PivotTables, advanced chart selection, interactivity with slicers, and visual storytelling. These are exactly the skills expected from modern data analysts, even before they move into Power BI, Python, or AI-driven analytics.

Finally, remember this: a good dashboard doesn’t just show data, it changes how people think. If this analysis helps educators guide students better or helps learners improve their own habits, then the dashboard has already delivered real value. That is the true goal of analytics.