Uni-variate data involves only one variable (feature/column) at a time.
Definition of Univariate Data
Univariate data is data that contains only one variable (one feature or one characteristic) collected from multiple observations.
👉 The word “uni” means one.
👉 So, univariate = one variable.
Simple Definition:
Univariate data is a type of data where analysis is done on a single variable without considering relationships with other variables.
2️⃣ What is a Variable?
A variable is any measurable characteristic that can take different values.
Examples of Variables:
Age
Height
Salary
Marks
Temperature
Gender
If we analyze only one of these at a time, it becomes univariate data.
3️⃣ Examples of Univariate Data
Example 1: Student Marks
Student Marks: A75 B82 C60 D90
✔ Only Marks is analyzed
✔ No comparison with other variables
➡ This is univariate numerical data
Example 2: Gender of Employees
Employee Gender 1.Male 2.Female 3.Male
✔ Only Gender
➡ This is univariate categorical data
Examples:
Age of customers
Salary of employees
Marks of students
Daily temperature
👉 No relationship with other variables is studied here.
2️⃣ What is Exploratory Data Analysis (EDA)?
EDA is the process of:
Understanding data
Summarizing data
Finding patterns, trends, and anomalies
Detecting outliers and errors
before applying machine learning or statistical models.
3️⃣ What is Uni-variate Graphical EDA?
Uni-variate Graphical EDA uses graphs and plots to visually analyze one variable.
Purpose:
✔ Understand data distribution
✔ Identify outliers
✔ Detect skewness
✔ Find data spread
✔ See frequency patterns
4️⃣ Why Use Graphical Methods?
Humans understand visuals faster than numbers
Easy to detect patterns & anomalies
Simplifies complex datasets
Essential first step in Data Science workflows
5️⃣ Types of Uni-variate Graphical EDA
Uni-variate graphical methods depend on data type: Data Type Common Graphs Categorical Bar Chart, Pie Chart Numerical Histogram, Box Plot, Density Plot
📌 A. Bar Chart (Categorical Data)
🔹 Definition:
A bar chart shows frequency or count of each category.
🔹 Example:
Gender = {Male, Female}
Department = {HR, IT, Sales}
🔹 Interpretation:
Height of bar → frequency
Taller bar → more observations
🔹 What We Learn:
✔ Most frequent category
✔ Least frequent category
✔ Class imbalance (important in ML)
🔹 Advantages:
Simple & clear
Best for discrete categories
🔹 Limitations:
Not suitable for continuous data
📌 B. Pie Chart (Categorical Data)
🔹 Definition:
Shows percentage contribution of each category.
🔹 Example:
Market share of companies
🔹 Interpretation:
Each slice represents proportion
Total = 100%
🔹 What We Learn:
✔ Relative proportion
✔ Contribution comparison
🔹 Limitations:
❌ Difficult with many categories
❌ Not good for precise comparison
👉 In Data Science, bar charts are preferred over pie charts.
📌 C. Histogram (Numerical Data)
🔹 Definition:
Histogram shows frequency distribution of numerical data using bins.
🔹 Example:
Marks of students
Salary distribution
🔹 Key Components:
X-axis → Value ranges (bins)
Y-axis → Frequency
🔹 What We Learn:
✔ Data distribution shape
✔ Skewness (Left / Right / Symmetric)
✔ Central tendency
✔ Presence of outliers
🔹** Types of Distribution:**
Normal (Bell-shaped)
Right-skewed (Positive skew)
Left-skewed (Negative skew)
Uniform
🔹 Importance in ML:
Many ML algorithms assume normal distribution.
📌 D. Box Plot (Numerical Data)
🔹 Definition:
Box plot summarizes data using five-number summary:
Minimum
Q1 (First Quartile)
Median
Q3 (Third Quartile)
Maximum
🔹 Visual Elements:
Box → IQR (Q3 - Q1)
Line inside box → Median
Dots outside → Outliers
🔹 What We Learn:
✔ Data spread
✔ Median position
✔ Outliers
✔ Skewness
🔹 Advantages:
Excellent for detecting outliers
Compact summary
🔹 Limitations:
Doesn’t show distribution shape clearly
📌 E. Density Plot (Numerical Data)
🔹 Definition:
Smooth curve showing probability density of data.
🔹 Difference from Histogram:
Histogram → bars
Density plot → smooth curve
🔹 What We Learn:
✔ Distribution shape
✔ Peaks (modes)
✔ Smooth visualization
🔹 Use Case:
Comparing distributions
Understanding continuous patterns
6️⃣ Skewness & Distribution Shape
Type Meaning Symmetric Mean ≈ Median-Right Skewed-mean > Median Left Skewed Mean < Median
👉 Important for feature transformation (log, sqrt).
7️⃣ Outliers in Uni-variate EDA
What are Outliers?
Extreme values that differ significantly from others.
Detected Using:
Box plot
Histogram
Why Important?
❗ Can distort:
Mean
Variance
ML model performance
8️⃣ Role in Data Science & ML Pipeline
Uni-variate Graphical EDA helps to:
✔ Decide data cleaning strategy
✔ Choose transformations
✔ Identify feature issues
✔ Improve model accuracy
9️⃣ Real-World Example
Dataset: Student Marks
Histogram → Understand score distribution
Box plot → Detect very low/high scores
Bar chart → Grade distribution
👉 Before applying prediction models.
🔟 Summary
Uni-variate Graphical EDA:
Focuses on one variable
Uses visual tools
Helps understand:
Distribution
Spread
Outliers
Skewness
Most Important Graphs:
✔ Bar Chart
✔ Histogram
✔ Box Plot
✔ Density Plot
Top comments (0)