Forem: Kelvin Kamau

From Messy Data to Business Decisions with Power BI

Kelvin Kamau — Mon, 09 Feb 2026 07:52:31 +0000

In many organizations, data exists everywhere but insights are often missing. Sales teams maintain spreadsheets, finance works with accounting systems, operations store data in separate tools, and management struggles to see the complete picture.

This is where analysts step in. Using tools like Power BI, analysts transform raw, messy data into meaningful insights that guide real business decisions. However, the real value does not come from dashboards alone — it comes from how data is cleaned, modeled, analyzed, and translated into action.

This article explains how analysts use Power BI skills, including data cleaning, modeling, DAX, and dashboard design, to create measurable business impact.

The Reality: Data Is Rarely Clean

In practice, business data is messy. Analysts often encounter problems such as:

Duplicate records
Missing values
Incorrect dates or prices
Different systems storing the same data differently
Manual data entry errors
Inconsistent naming conventions

For example, a sales dataset may contain:

Customers entered with multiple spellings
Negative prices caused by system errors
Missing sales representatives
Orders recorded in different formats

Before analysis even begins, analysts must clean and standardize data. In Power BI, this is done using Power Query, where analysts:

Remove duplicates
Fix data types
Replace missing values
Standardize fields
Merge datasets from multiple sources

Without proper cleaning, dashboards can produce misleading results, leading to poor decisions.

Structuring Data for Analysis

Once cleaned, data must be structured properly. Analysts build data models using fact and dimension tables connected through relationships.

For example:

A sales fact table stores transaction data.
Dimension tables store products, customers, dates, and regions.

This structure allows questions such as:

Which region generates the most profit?
Which products are underperforming?
Which customers drive revenue growth?

Good data modeling ensures reports calculate totals correctly and respond quickly to filters.

DAX: Turning Data into Business Metrics

After structuring data, analysts use DAX (Data Analysis Expressions) to create business metrics.

DAX transforms raw numbers into meaningful performance indicators.

Examples include:

Total Revenue
Gross Profit
Profit Margin
Year-to-Date Sales
Growth percentages
Customer retention rates

For instance, a company may track revenue, but management actually needs to know:

Are profits increasing?
Which regions are declining?
Are discounts affecting margins?

DAX allows analysts to create measures that directly answer these business questions.

Dashboards: Communicating Insights Clearly

Dashboards translate analysis into visuals decision-makers can quickly understand.

Effective dashboards:

Highlight key performance indicators
Show trends over time
Compare regions or products
Reveal risks and opportunities

For example, a Power BI dashboard may reveal:

Sales growing but profit declining
A region losing customers
Certain products driving most revenue
Delivery delays affecting performance

The analyst’s role is to design dashboards that tell a story rather than overwhelm users with charts.

From Dashboards to Decisions

The real value of Power BI appears when dashboards influence decisions.

Examples of business impact include:

Marketing reallocates budget after discovering profitable customer segments.
Sales teams focus on high-performing products.
Management identifies regions needing operational improvement.
Pricing strategies change after discount impacts become visible.

A well-designed dashboard moves conversations from guessing to evidence-based decision-making.

The Analyst’s True Role

An analyst’s job is not simply to build visuals. It involves:

Understanding business problems
Cleaning and structuring data
Creating accurate metrics
Designing understandable reports
Explaining insights to stakeholders
Supporting decision-making

Technical skills are important, but communication and business understanding are equally critical.

Common Challenges Analysts Solve

Power BI analysts frequently address questions such as:

Why is revenue increasing but profits are not?
Which customers are most valuable?
Where are operational bottlenecks occurring?
Which products should be discontinued?
How do seasonal trends affect performance?

Answering these questions requires combining technical skills with business reasoning.

Why Power BI Is Effective for Business Analytics

Power BI supports analysts by allowing them to:

Connect to multiple data sources
Clean and transform data
Build efficient data models
Create powerful calculations using DAX
Design interactive dashboards
Share insights across organizations

This combination makes Power BI a strong platform for turning data into actionable intelligence.

Conclusion

Messy data alone does not drive business success. Impact comes from analysts who transform that data into insights using modeling, DAX, and dashboards.

Power BI enables analysts to move organizations from reactive decisions to proactive strategies. By cleaning data, structuring information, creating meaningful measures, and presenting insights clearly, analysts help businesses act confidently and competitively.

In the end, dashboards do not change organizations, informed decisions do. Analysts bridge that gap.

Understanding Schemas and Data Modelling in Power BI

Kelvin Kamau — Mon, 02 Feb 2026 13:10:00 +0000

Power BI is widely used for creating business reports and dashboards, but the quality of insights produced depends heavily on how data is structured before visualization. This structure is known as data modelling, and it plays a critical role in performance, usability, and accuracy of reports.

In this article, we explore schemas and data modelling concepts in Power BI, including fact tables, dimension tables, relationships, star schema, and snowflake schema, and explain why proper modelling is essential for reliable analytics.

What is Data Modelling in Power BI?

Data modelling in Power BI refers to organizing data tables and defining relationships between them so analysis and reporting become efficient and accurate.

Instead of placing all data in one massive table, data modelling separates information into logical tables connected through relationships. This structure makes data easier to analyze and prevents errors such as double counting or incorrect aggregations.

A well-designed model allows users to create reports without manually joining or cleaning data repeatedly.

Fact Tables

A fact table contains measurable business events or transactions. These are numeric values that analysts aggregate to produce insights.

Common examples include:

Sales transactions
Revenue amounts
Order quantities
Costs and profits
Website visits or transactions

A fact table usually contains:

Numeric measures such as sales, cost, or quantity
Foreign keys linking to dimension tables
Transaction-level data

Example fields in a sales fact table:

OrderDate	ProductID	CustomerID	SalesAmount	Quantity
2024-01-05	P01	C002	500	2

Fact tables tend to be large because they store transactional data.

Dimension Tables

A dimension table provides descriptive information used to filter, group, or categorize facts.

Dimensions answer questions such as:

Which product was sold?
Which customer bought it?
In which region did sales occur?
Which salesperson handled the order?

Example dimension tables include:

Product dimension
Customer dimension
Date dimension
Salesperson dimension
Region dimension

Example Product dimension:

ProductID	ProductName	Category	Brand
P01	Laptop Pro	Electronics	TechBrand

Dimension tables are usually smaller and contain descriptive attributes.

Relationships in Power BI

Relationships connect fact tables to dimension tables. Power BI uses these relationships to filter and aggregate data correctly.

Most relationships follow a one-to-many structure:

One product → many sales records
One customer → many transactions
One region → many orders

In Power BI, relationships allow filters applied to dimensions (like selecting a country) to automatically affect fact data.

Incorrect relationships often lead to wrong totals or missing values in reports.

Star Schema

The star schema is the recommended modelling approach in Power BI.

In a star schema:

A central fact table connects directly to several dimension tables.
Dimensions are not connected to each other.
The structure visually resembles a star.

Example structure:

           Product
               |
Customer — Sales Fact — Date
               |
            Region

Advantages of Star Schema

Fast query performance
Easy to understand
Simplifies report building
Reduces modelling complexity
Improves aggregation accuracy

Because Power BI’s engine is optimized for star schemas, reports built on this model usually perform better.

Snowflake Schema

A snowflake schema is similar to a star schema but dimensions are further normalized into multiple related tables.

Example:

Sales Fact → Product → Product Category

Instead of keeping all product details in one dimension, category data is stored separately.

Advantages

Reduces data redundancy
Saves storage space

Disadvantages

More complex relationships
Harder to maintain
Slower performance in Power BI
Confusing for report users

For Power BI, snowflake schemas are usually discouraged unless necessary.

Why Good Data Modelling Matters

Good modelling directly impacts report quality and performance.

Faster Reports

Proper schemas allow Power BI to process queries efficiently, reducing report load time.

Accurate Calculations

Poor models often lead to duplicated counts or incorrect totals. Correct relationships prevent these issues.

Easier Report Building

A clean model lets users drag and drop fields easily without worrying about complex joins.

Scalable Analytics

Well-structured models allow future data additions without breaking reports.

Better User Experience

Users interact with clean dimensions rather than messy raw data tables.

Best Practices for Power BI Data Models

Practical modelling guidelines include:

Use a star schema whenever possible.
Keep fact and dimension tables separate.
Avoid many-to-many relationships unless necessary.
Use clear naming conventions.
Remove unnecessary columns.
Maintain clean date tables.

Conclusion

Schemas and data modelling form the foundation of successful Power BI reports. Understanding fact tables, dimension tables, relationships, and schema types helps analysts build models that are both efficient and accurate.

While both star and snowflake schemas organize data effectively, Power BI performs best with star schemas due to simplicity and speed.

Investing time in proper data modelling ensures faster dashboards, accurate insights, and better decision-making across organizations.

Using Microsoft Excel for Basic Data Analysis (Beginner-Friendly Guide)

Kelvin Kamau — Sun, 25 Jan 2026 09:20:51 +0000

Microsoft Excel is one of the easiest tools to start learning data analysis. Even without coding, Excel can help you organize data, summarize it, and spot patterns that support decision-making.

In this article, I will use an HR dataset to demonstrate how Excel can be used for basic data analysis using simple, beginner-friendly steps.

1. Understanding the Dataset

Before doing any analysis, it is important to understand what the dataset contains.

In this HR dataset:

Each row represents one employee
Each column represents a type of employee information

Some of the columns included in the dataset are:

Employee ID
First Name
Last Name
Department
Salary
Hire Date
Age
Gender
Performance Score
Employee Type
Office Location
Remote Work Status

This dataset can help answer common HR questions such as:

Which department has the most employees?
What is the average salary in the company?
How many interns are based in Nairobi?
Which employees have low performance scores?

2. Converting the Data into an Excel Table

Excel Tables make analysis easier because they support:

quick sorting and filtering
automatic formatting
structured references in formulas
easier PivotTable creation

To convert the dataset into a table:

Click anywhere inside the dataset
Press Ctrl + A to select the dataset
Press Ctrl + T to create a table
Tick My table has headers
Click OK

Screenshot 2: Create Table Dialog

After creating a table, dropdown arrows appear on the column headers. These dropdowns allow sorting and filtering.

3. Sorting Data (Example: Highest Salary)

Sorting helps you arrange your data in order so that you can quickly identify high or low values.

For example, sorting the Salary column can help you find:

the highest-paid employees
the lowest-paid employees
salary patterns across departments

To sort salary from highest to lowest:

Click the dropdown arrow on the Salary column
Select Sort Largest to Smallest

This makes it easy to identify the top earners and understand salary differences across employees.

4. Filtering Data (Example: Interns in Nairobi)

Filtering allows you to display only the records you want to focus on without deleting anything.

A good example is finding employees who are:

Interns
and based in Nairobi

To filter interns in Nairobi:

Click the filter dropdown on Employee Type and select Intern
Click the filter dropdown on Office Location and select Nairobi

Filtering helps you narrow down a large dataset to a specific group, which is useful for reporting and decision-making.

5. Using Formulas for Basic Analysis

Excel formulas help you calculate results quickly and accurately. Instead of counting manually, Excel can summarize your dataset in seconds.

5.1 Counting Employees Using COUNTIFS

If you want to count employees who meet more than one condition, Excel provides the COUNTIFS function. This is useful when you are filtering employees by category and location.

For example, to count how many interns are based in Nairobi, you can use:

 =COUNTIFS(N2:N877,"Intern",O2:O877,"Nairobi")

This gives you the total number of interns in Nairobi instantly, which is useful for reporting and workforce planning.

5.2 Finding the Average Salary Using AVERAGE

To understand salary levels in the organization, you may want to calculate the average salary across all employees. Excel provides the AVERAGE function for this.

You can calculate the average salary using:

=AVERAGE(E2:E877)

This gives a quick summary of the company’s general salary level and can be used for benchmarking and planning.

5.3 Finding the Highest and Lowest Salary Using MAX and MIN

Sometimes, it is helpful to know the salary range in the dataset. Excel provides the MAX and MIN functions to find the highest and lowest values.

To find the highest salary:

 =MAX(E2:E877)

To find the lowest salary:

 =MIN(E2:E877)

These formulas help you identify the salary range and quickly spot unusually high or low values.

6. Summarizing Data with PivotTables

PivotTables are one of Excel’s most useful tools for data analysis because they allow you to summarize large datasets without writing many formulas.

A simple example is counting how many employees are in each department.

To create a PivotTable:

Click anywhere inside the table
Go to Insert → PivotTable
Choose New Worksheet
Drag Department into Rows
Drag Employee ID into Values and ensure it is set to Count

This summary helps you quickly identify which departments have the highest number of employees.

7. Visualizing Results Using Charts

Charts make it easier to understand results because they present the information visually. Instead of reading many numbers, you can see patterns at a glance.

A simple chart can be created from the PivotTable results to show employee count by department.

To create a chart:

Click the PivotTable results
Go to Insert → Column Chart

This chart makes it easier to compare departments and identify which ones have more or fewer employees.

8. Using Conditional Formatting to Spot Patterns

Conditional formatting allows Excel to automatically highlight values based on rules. This helps you notice patterns such as low scores, high values, or unusual results without scanning every row manually.

For example, you can apply conditional formatting to the Performance Score column to quickly identify low and high performers.

To apply conditional formatting:

Select the Performance Score column
Go to Home → Conditional Formatting
Choose Color Scales or Highlight Cells Rules

This makes it easier to identify performance patterns across employees.

Conclusion

Microsoft Excel is a powerful tool for basic data analysis because it allows you to:

structure data properly using tables
sort and filter information quickly
calculate summaries using formulas
create PivotTables for fast reporting
visualize results using charts
highlight trends using conditional formatting

With these basic features, you can turn raw employee records into meaningful insights that support decision-making in an organization.

Getting Started with Git: Track Your Code and Use GitHub with Confidence

Kelvin Kamau — Sun, 18 Jan 2026 10:06:28 +0000

If you're learning to code, you’ll quickly notice something: writing code is only half the work. The other half is managing change.

You’ll build a feature, adjust it, break something, fix it, then later wish you could “go back to the version that worked yesterday.”

That’s exactly the problem Git was built to solve.

In this article, you’ll learn what Git is, why version control matters, and how to use Git Bash + GitHub to push and pull code confidently — plus how to track changes properly.

What Git is (and why version control matters)

Git is a version control system. It helps you track changes made to your files over time.

Instead of saving multiple copies like:

project_final.zip
project_final_real.zip
project_final_please_work.zip

Git stores a clean history of your project, and you can move between versions whenever you need to.

Version control is important because it helps you:

recover from mistakes
see what changed and when
collaborate without overwriting each other’s work
experiment safely

If Git is the tool that tracks changes locally on your computer, GitHub is the platform that stores Git projects online. It’s where you back up your work and share it with others.

Installing Git Bash (Windows)

On Windows, the easiest way to use Git is through Git Bash — a terminal that lets you run Git commands.

To install it:

Download Git from the official site: https://git-scm.com/downloads
Install it (default settings are fine)
Search for Git Bash in the Start Menu and open it

Once Git Bash opens successfully, you’re ready to start using Git.

Connecting Git to your GitHub account

Before you start saving versions of your work, Git needs to know who you are. This information is attached to your commits (think of commits like signed checkpoints).

In Git Bash, set your name and email:

git config --global user.name "Your Name"
git config --global user.email "your-email@example.com"

You can confirm your settings with:

git config --global --list

Tip: Use the same email address that you use on GitHub so your commits show up correctly on your profile.

Your first Git workflow: tracking changes locally

A Git project starts inside a folder. Once Git is enabled, it begins watching your files and tracking what changes over time.

If you already have a project folder, move into it. Then initialize Git:

git init

From this point, Git starts monitoring the folder.

Now here’s where beginners often get confused: Git has two main stages before your work is saved in history.

1) Staging

Staging means selecting what you want to include in your next snapshot.

git add .

2) Committing

A commit is the snapshot itself — a saved point in history.

git commit -m "Initial commit"

If you ever want to check what Git sees in your project, use:

git status

This command tells you:

which files changed
which ones are staged
which ones are still untracked

Pushing your code to GitHub

Saving your work locally is great, but pushing to GitHub is what makes your project shareable and safely backed up online.

To push code, you’ll need a GitHub repository first:

Log in to GitHub
Create a new repository
Copy the repository link (HTTPS is fine for beginners)

Back in Git Bash, connect your local project to GitHub:

git remote add origin https://github.com/USERNAME/REPOSITORY.git

Then push your commits:

git branch -M main
git push -u origin main

Once that completes, refresh your GitHub repository page — you should see your files there.

That’s your first successful push 🎉

Pulling code from GitHub (keeping your project updated)

As soon as you start working with GitHub, your project now has two versions:

your local version (on your computer)
the remote version (on GitHub)

When you want to download the latest updates from GitHub into your computer, you pull:

git pull origin main

This is especially important when:

you’re collaborating with teammates
you edited the repo on GitHub directly
you’re switching between devices

If you don’t have the project locally at all (first time), you use clone:

git clone https://github.com/USERNAME/REPOSITORY.git

Cloning is like copying the whole project from GitHub to your machine, including the Git history.

Tracking changes like a developer

Git becomes truly powerful when you use it to understand what’s happening in your project, not just to upload code.

Here are the most useful commands for tracking changes:

Checking what changed

git status

This is your project dashboard.

Seeing the exact line changes

git diff

This shows what changed line-by-line before you commit.

Viewing your project history

git log --oneline

This gives you a clean summary of your commits and messages.

The simple daily workflow

Most developers repeat the same cycle:

git status
git add .
git commit -m "Describe what changed"
git push

And when working with others:

git pull

If you master these commands, you can confidently manage almost any beginner project.

Final thoughts

Git is not just a tool you must learn — it’s a safety net.

It helps you build projects without fear, recover quickly from mistakes, and work like a professional even on small personal projects.

Once you get comfortable with commits, push, pull, and tracking changes, GitHub becomes a natural part of your workflow.

Happy coding 🚀