Data Normalization Using SQL, Power BI, Python and Excel

Imagine you have a messy box of toys. Some toys are repeated, some are broken, and some are all mixed up. You want to organize them so that everything is easy to find and use.

That’s exactly what data normalization does for information.

Data Normalization:

Normalization means:

• Cleaning data

• Reducing duplicates

• Making information consistent

• Splitting big messy tables into smaller, neat ones.

  
  

Think of data normalization like organizing your home.

• SQL → putting toys in separate labeled boxes

• Power BI → lining toys up so each row is one toy

• Python → either organizing the toys or scaling them so they fit nicely on the shelf

• Excel → cleaning, separating, and aligning toys neatly on shelves

  
  

 Normalization in SQL:

SQL stores data in tables. Instead of repeating information, we separate it into related tables.

Example:

Messy table:

Student	Grade	Subject	Teacher
Alice	10	Math	Mr. Smith
Alice	10	English	Mrs. Johnson
Bob	11	Math	Mr. Smith

Step 1: Split into tables

CREATE TABLE Students (

    StudentID INT PRIMARY KEY,

    Name VARCHAR(50),

    Grade INT

);

CREATE TABLE Subjects (

    SubjectID INT PRIMARY KEY,

    SubjectName VARCHAR(50),

    Teacher VARCHAR(50)

);

CREATE TABLE Enrollment (

    EnrollmentID INT PRIMARY KEY,

    StudentID INT,

    SubjectID INT,

    FOREIGN KEY (StudentID) REFERENCES Students(StudentID),

    FOREIGN KEY (SubjectID) REFERENCES Subjects(SubjectID)

);

·       Each piece of info is stored once.

·       Easy to update: If Mr. Smith changes, you only change it in one place.

2. Normalization in Power BI

Normalization in Power BI means each row should describe just ONE thing.

• One row = one event, one transaction, or one record

• No combining multiple values in the same row.

Messy table:

Student	Math	English
Alice	95	88
Bob	80	90

Step: Unpivot columns

Student	Subject	Score
Alice	Math	95
Alice	English	88
Bob	Math	80
Bob	English	90

·       Each row is one piece of information

·       Easier to build charts, graphs, and calculations

3. Normalization in Python

Python has two main ways to normalize:

a) Organize data (like SQL)

import pandas as pd

data = {

    'Student': ['Alice', 'Alice', 'Bob'],

    'Subject': ['Math', 'English', 'Math'],

    'Teacher': ['Mr. Smith', 'Mrs. Johnson', 'Mr. Smith']

}

df = pd.DataFrame(data)

students = df[['Student']].drop_duplicates()

subjects = df[['Subject', 'Teacher']].drop_duplicates()

·       Reduces duplicates, separates info into clean tables

b) Scale numeric data (Min-Max normalization)

from sklearn.preprocessing import MinMaxScaler

import pandas as pd

data = {'Score': [80, 90, 95]}

df = pd.DataFrame(data)

scaler = MinMaxScaler()

df['Score_Normalized'] = scaler.fit_transform(df[['Score']])

print(df)

Output:

Score	Score_Normalized
80	0.0
90	0.5
95	1.0

·       All numbers are now on the same scale (0 to 1)

·       Useful for graphs, comparisons, and machine learning

4.Normalization in Excel (Most Common Tool!)

Excel normalization is mostly about cleaning and structuring data properly.

Student	Subjects	Scores
Alice	Math, English	95, 88
Bob	Math	80

Problems:

• Multiple values in one cell

• Hard to analyze

• Charts won’t work properly

  Normalized excel table:

Student	Subject	Score
Alice	Math	95
Alice	English	88
Bob	Math	80

 In Excel, normalization is mostly about cleaning and restructuring data so it’s easy to analyze. You start by removing duplicates using Data → Remove Duplicates to ensure the same information isn’t repeated again and again. When multiple values are packed into one cell, you can split columns using Text to Columns so each value gets its own place. For more advanced cleaning, Power Query helps by unpivoting columns, which turns wide, messy tables into clean, row-based data where each row represents just one record. Finally, you can separate data into different tables like Students, Subjects, and Scores, instead of keeping everything in one big sheet. When Excel data is normalized this way, each row tells a single clear story, formulas become simpler and more reliable, and Pivot Tables work faster and give much better insights.

Tool	How Normalization Works	Why It Helps
SQL	Split into tables + keys	No duplicates, easy updates
Power BI	Unpivot & structure rows	Clean visuals, better measures
Python	Split tables or scale numbers	Consistency, ML-ready data
Excel	Clean, unpivot, remove duplicates	Easy analysis & reporting

Data normalization simply means keeping your data tidy, clean, and well organized. When data is organized properly, it avoids repeated information, reduces mistakes, and saves storage space. Clean and structured data also makes analysis, visualization, and reporting much faster and easier. Whether you use SQL, Power BI, Python, or Excel, the idea of normalization stays the same—only the way you apply it changes from tool to tool.

Think of normalization like organizing your toys:

• One toy per box → SQL (each table has a clear purpose)

• Line them up neatly → Power BI (one row equals one record for easy visuals)

• Resize or group them for easy comparison → Python (scaling and structuring data)

• Clean, sort, and arrange toys on shelves → Excel (remove duplicates, split columns, and use Power Query to keep data tidy)

No matter the tool, the goal is always the same: make data easy to understand, easy to manage, and easy to use.

• 

  Image Source:Google