Forem: Mathew Leshan

Using PowerBI toTranslate Messy Data DAX and Dashboards-into Actionable Insights

Mathew Leshan — Mon, 09 Feb 2026 16:25:22 +0000

Raw data is loud.

It’s full of missing values, strange formats, duplicated rows, and numbers that don’t seem to agree with each other. On its own, raw data doesn’t tell a story..

This is where analysts come in.

Using Power BI, analysts act as translators turning messy data, complex DAX calculations, and dashboards into clear, actionable decisions.

This article walks through how that translation actually happens in the real world.

Step 1: Starting with Messy Data

Let’s be honest, clean data is rare.

Typical datasets (hospital records, sales logs, crop yields, surveys) usually come with:

Null or missing dates
Inconsistent naming (Nairobi, NRB, Nairobi County)
Mixed data types
Duplicates
Columns are doing too many jobs

Power BI doesn’t magically fix this.

Power Query is an inbuilt softwarein PowerBI where Analysts Clean and standardise data

Power Query is the first layer of translation.

What Analysts Do Here:

Remove duplicates
Replace or flag null values
Standardize text and date formats
Split columns into usable fields
Filter out irrelevant records

Messy CSV
↓
Power Query
↓
Structured, trusted tables

Step 2: Modeling the Data Like the Real World Works

Once data is clean, analysts don’t jump straight to visuals.

They model relationships.

Example:

A patient can have many visits
A farmer can grow multiple crops
A customer can make many purchases

One → Many relationships

Why This Matters:

Prevents incorrect totals
Ensures filters work properly
Makes calculations accurate
Reflects real-world logic

Bad relationships = misleading insights.

Step 3: DAX

DAX (Data Analysis Expressions) is where Power BI becomes powerful.

Not because it’s complex — but because it’s context-aware.

Analysts Use DAX to Answer Questions Like:

What is the average cost per visit, not just total cost?
How do yields change over time?
What happens when we filter by region, date, or category?

Measures vs Columns (A Critical Distinction)

Calculated Columns	Measures
Row-level	Aggregated
Stored in table	Calculated on demand
Heavy	Efficient

Analysts favor measures because:

They respond to filters
They keep models lean
They reflect real-time context

Iterator Functions: Thinking Row by Row

Functions like:

SUMX
AVERAGEX
MINX
MAXX

Example:
DAX
AVERAGEX(Visits, Visits[Cost] * Visits[Discount])
This is how analysts move from simple totals to business logic.

Step 4: Dashboards That Answer Questions

A good dashboard answers:

What’s happening?

Why is it happening?

What should we do next?

A bad dashboard just shows everything.

What Analysts Focus On:
Clear KPIs

Trends over time

Comparisons (before vs after)

Interactive filters (slicers)

Minimal but meaningful visuals

Data → Insight → Decision
If a visual doesn’t support a decision, it doesn’t belong there.

Turning Dashboards into Action

This is the most important part and the most misunderstood.

Example Actions:
A hospital reallocates staff based on patient load

A county adjusts crop support based on yield trends

A business cuts costs after identifying inefficiencies

Power BI doesn’t make decisions.

People do using insights Power BI reveals.

Common Translation Mistakes Analysts do;
-Cleaning data in DAX instead of Power Query

Ignoring data relationships
Overusing calculated columns
Building dashboards with no clear question
Treating Power BI like Excel with better charts

Analysts:

Clean it

Structure it

Calculate it

Visualize it

Explain it

Power BI is just the tool.

The real value lies in the translation — turning chaos into clarity, and clarity into action.

If you can do that, you’re not just building dashboards.

Your driving decisions.

If you’re learning Power BI:

Master Power Query first

Understand relationships deeply

Learn DAX with context, not memorisation

Design dashboards with intent

Because the goal isn’t prettier charts, but better decision-making.

Power BI Data Modelling Demystified ⭐

Mathew Leshan — Wed, 04 Feb 2026 05:52:33 +0000

Introduction: Where Dashboards Go to Live… or Die

You can clean your data perfectly.

You can build flashy visuals.

You can even write clever DAX.

And yet… your Power BI report is slow, confusing, or worse — wrong.

Why?

Because behind every great Power BI dashboard is an unsung hero (or villain):

👉 the data model.

Think of data modelling as the plumbing of Power BI. When it’s done right, everything flows smoothly. When it’s done badly, nothing works the way you expect. This article breaks down Power BI data modelling in plain language, covering schemas, fact and dimension tables, relationships, and why good modelling is critical for performance and accurate reporting — especially when working with real-world datasets like hospital records or Kenya Crops data.

What Is Data Modelling in Power BI?

Data modelling is how you structure tables and define how they connect inside Power BI.

A good model determines:

How filters move across tables
How measures calculate values
How fast your report runs
Whether your numbers can be trusted

Fact Tables vs Dimension Tables

At the heart of Power BI modelling are two table types.

Fact Tables 📊

Fact tables store events and measurements things you want to analyse.

Examples:

Hospital admissions
Crop harvest volumes
Patient visits
Sales transactions

Typical characteristics:

Many rows
Numeric values (counts, totals, averages)
Foreign keys linking to dimensions

FACT_Admissions
AdmissionID
PatientID
DepartmentID
DateID
LengthOfStay
Cost

Dimension Tables 🧭

Dimension tables add context and meaning to facts.

Examples:

Date
Department
Crop type
Region

Typical characteristics:

Fewer rows
Descriptive columns
One unique key per row

DIM_Department
DepartmentID
DepartmentName
HospitalWing

Star Schema ⭐ (Power BI’s Best Friend)

The star schema is the gold standard for Power BI.

Structure

One central fact table
Multiple dimension tables
All dimensions connect directly to the fact table
```
DIM_Date
   |
```
DIM_Patient — FACT_Admissions — DIM_Department
|
DIM_Doctor

Why Star Schema Works So Well

Faster performance
Simple relationships
Cleaner filter flow
Easier DAX calculations
Easy to understand and maintain

Snowflake Schema ❄️ (Looks Fancy, Works Harder)

Snowflake schema is a normalised version of the star schema.

Structure

Dimension tables are split into multiple related tables
Creates extra joins and relationship chains

FACT_Admissions
|
DIM_Department
|
DIM_HospitalWing

Pros

Reduces data duplication
Can save storage

Cons (in Power BI)

Slower performance
More complex relationships
Harder DAX
Filters can behave unexpectedly

📌
Snowflake schemas belong in databases — star schemas belong in Power BI.

Relationships in Power BI 🔗

Relationships define how tables talk to each other.

Best Practice Relationship Setup

One-to-many (1:*)
Many to one
Many to many

What Can Go Wrong?

Bad relationships can:

Inflate totals
Break slicers
Cause ambiguous paths
Slow down reports

Why Good Data Modelling Is Critical 🚨

1. Performance

Star schemas reduce joins and memory usage, making reports faster.

2. Accuracy

Correct relationships ensure filters behave logically.

3. Simpler DAX

Good models reduce the need for complex CALCULATE and FILTER logic.

4. Scalability

Adding new KPIs or dimensions becomes easy instead of painful.

5. Trust

Decision-makers rely on reports — wrong numbers destroy confidence.

Power BI Data Modelling Best Practices ✅

Use star schema whenever possible
Keep fact tables narrow
Create a proper Date table
Use numeric surrogate keys
Avoid many-to-many relationships
Avoid bi-directional filters
Clean data in Power Query, not DAX

Conclusion: Model First, Visualise Second 🎯

Data modelling is not optional in Power BI — it’s foundational. A well-designed model ensures performance, accuracy, and clarity, turning raw datasets into reliable insights.

Before adding visuals or writing DAX, always ask:

Does my model make sense?

Because in Power BI, great dashboards don’t start with charts —

They start with the model.

📊 Excel-ently Done: Turning Everyday Numbers into Powerful Insights with Microsoft Excel

Mathew Leshan — Sun, 25 Jan 2026 20:58:17 +0000

Introduction to Excel

Let’s be honest, when most people hear “Microsoft Excel,” they imagine endless rows, tiny boxes, and numbers that make their eyes glaze over. Not very exciting, right?

But what if I told you that behind those neat little cells lies a powerful data analytics tool capable of turning raw, messy numbers into clear stories and smart decisions?

Whether you’re analysing sales, tracking performance, or just trying to make sense of data without writing a single line of code, Excel is often the first best friend of every data analyst. In this article, we’ll explore how Microsoft Excel can be used for basic data analysis, using simple language that even a complete beginner can understand.

What is Microsoft Excel?

Microsoft Excel is a spreadsheet application used to store, organise, analyse, and visualise data. It works by arranging data into rows and columns, making it easy to read, edit, and analyse information.
Excel

Excel is widely used because:

It is easy to learn

It requires no programming knowledge

It is available in many organisations

It handles both small and large datasets

For beginners in data analytics, Excel is usually the starting point before moving on to advanced tools such as SQL, Power BI, or Python.

Understanding Rows, Columns, and Cells

Excel data is organised in a grid format:

Rows run horizontally and are numbered (1, 2, 3…)

Columns run vertically and are labelled with letters (A, B, C…)

A cell is where a row and column meet (e.g., A1)

Each cell can contain text, numbers, dates, or formulas. This structure makes Excel ideal for organising datasets such as student records, sales reports, or survey results.

Entering and Cleaning Data in Excel

Before analysing data, it must be clean and well-organised. Dirty data can lead to incorrect results.

Excel provides several tools for data cleaning, such as:

Remove Duplicates to eliminate repeated entries

Find and Replace to correct spelling or formatting errors

Formatting options to standardise dates and numbers

These features help ensure that the data is accurate and ready for analysis.

Using Basic Excel Formulas for Analysis

Addition: =A1+B1
Subtraction: =A1-B1
Multiplication: =A1*B1
Division: =A1/B1
Percentage: =A1*10%
Sum a Range: =SUM(A1:A10)
Average a Range: =AVERAGE(A1:A10)
Count Cells: =COUNT(A1:A10) (counts numbers)
Maximum: =MAX(A1:A10)
Minimum: =MIN(A1:A10)

One of Excel’s greatest strengths is its ability to perform automatic calculations using formulas.

Some basic formulas used in data analysis include:

SUM – adds values

=SUM(B2:B10)

AVERAGE – finds the mean

=AVERAGE(B2:B10)

COUNT – counts numeric entries

IF – applies logical conditions

=IF(C2>=50,"Pass","Fail")

These formulas allow users to analyse data quickly and efficiently without manual calculations.

Sorting and Filtering Data

When working with large datasets, sorting and filtering become essential.

Sorting arranges data in ascending or descending order (e.g., highest to lowest scores)

Filtering displays only selected data based on conditions (e.g., scores above 70)

This helps users focus on specific information and identify patterns easily.

Visualising Data with Charts

Numbers alone can be hard to interpret. Charts help turn data into visual insights.

Excel allows users to create:

Bar charts

Line charts

Pie charts

Column charts

For example, a pie chart can show percentage distribution, while a bar chart can compare values across categories.

Why Excel is a Must-Have Skill for Data Analysts

Excel remains relevant in data analytics because:

It is widely used across industries

It supports quick analysis and reporting

It builds a strong foundation for advanced analytics tools

It helps develop analytical thinking

For beginners, mastering Excel is like learning the alphabet of data analytics.

Conclusion: Small Cells, Big Power 💡

Microsoft Excel may look simple, but it is a powerful tool for basic data analysis. From organising and cleaning data to performing calculations and creating charts, Excel helps transform raw data into meaningful insights. For anyone starting their data analytics journey, Excel is not just usefult’s essential.

Delving into data science

Mathew Leshan — Sat, 17 Jan 2026 05:02:08 +0000

Understanding Git, GitHub, push, pull and version control

This article will help you understand Git and GitHub, terms like push, pull, tracking changes, and version control

Git Bash is a tool that allows you to interact with Git using simple commands. It is basically a channel through which you can communicate with GitHub.

To install your Git Bash, visit GITBASH. Once installed, you can open Git Bash from your application and you will be directed to a black window where you will begin.

Connecting Git Bash to GitHub

Now, after creating your account on GitHub, you have to configure Gitbash to git hub by inputting commands that will generate a key which you'll later input to GitHub and connect them.

Pushing Code to GitHub

Pushing code means making changes to a project and sending those changes to GitHub.This step uploads your work to the online repository, making it available as a backup or for collaboration.

After pushing, your code is safely stored online. If your computer fails or you switch devices, your work remains accessible on GitHub.

Pulling Code from GitHub

Pulling code is the opposite of pushing. It means downloading the latest version of a project from GitHub to your computer.

Pulling is useful when:

You are working on multiple devices

Someone else has updated the project

Tracking Changes in Git

Tracking changes means checking what files have been modified, added, or deleted since the last save. Git continuously watches your project and notes every difference
This helps you to stay up to date with your projects, especially if you are working on different devices or collaborating on a project.

Version control using Git

Version control is basically a system that records every change made to a project over time. Instead of saving multiple copies of the same file with different names, Git automatically keeps a history of all changes.

This means you can:

Go back to an earlier version of your work

See exactly what changed and when

Work safely without fear of losing progress

Version control is very important when working in a group since it helps not to overwrite each others work.