ETX2250/ETF5922

Data Visualisations in Power BI

Lecturer: Kate Saunders

Department of Econometrics and Business Statistics


  • etx2250-etf5922.caulfield-x@monash.edu
  • Lecture 3
  • <a href=“dvac.ss.numbat.space”>dvac.ss.numbat.space


Learning Objectives

Learning Objectives

  • Understand the importance of data visualisation

  • Become familiar with Power BI’s visualisation capabilities

  • Learn how to use Power BI’s visualisation pane

  • Learn best practices for effective data visualisation

Power BI

What is Power BI?

What is Power BI?

It is a common tool used by business analysts.

Does the standard stuff:

  • Data Integration: Importing data and connecting to multiple data sources (Excel, databases, cloud services).
  • Data Transformation: Can use Power Query to clean and transform data.
  • Data Visualisation: Create charts, graphs, and maps (also interactive features).
  • Collaboration and Sharing: Publish reports and dashboards for sharing with others.

Why Use Power BI?

What are the key advantages?

  • Ease of Use: Drag-and-drop interface suitable for beginners.
  • Employers Use It: Very common tool for visualisation in business analytics.
  • Reporting: It’s highly useful for automatic and repeated reporting.
  • Real-Time Analytics: Allows real-time data updates for timely insights.

Other choices

There are many options for business analysts

Data visualisation tools (not particular order):

And many more …

We can’t teach them all, so we aim to give you transferable foundations.

Power BI vs R

Power BI

Power BI is ideal for day-to-day business reporting.

Pros:

  • User-friendly (no programming)
  • Easy to create dashboards
  • Real-time capabilities, and
  • Straightforward sharing

Cons:

  • Limited Customisation
  • Cost

R

R is great for data scientists and analysts who need in-depth control over visualisations.

Pros:

  • Highly customisable
  • Suitable for advanced analytics
  • Open source (aka. free)
  • Automation / reproducibility (more than plug and play)

Cons:

  • Higher hurdle to entry - need to code.
  • Simple tasks don’t need customisation

Power BI and R

Important

  • If you master data visualisation in both Power BI and R, you’ll have a diverse skillset.

  • We exposure you to these tools, so you can see the full spectrum of data visualisation tools used in business analytics.

  • We know R is harder, but its is important to challenge yourselves!

Let’s look at some examples

Industry Applications of Power BI

There are many, many examples!


  • Tourism Australia Which country has the most flights each week into Melbourne Airport? NZ

  • Vic Health What was the most common infection in the Local Government area of Monash? Influenza

  • Recycling Victoria How many tonnes of paper were collected from Monash City Council in 2022? 52% 5860t

Notice these examples show mostly simple visualisations (e.g. map, line plot, bar charts) linked with tables. This is where Power BI excels.

Your turn

Your turn

  • Take some time to explore these dashboards

  • Consider the ease of use and functionality

  • Form an opinion on whether these are good or bad plots

    • Think about graphical principles, visual hierarchy, gestalt principles etc

  • Try answering my questions from the previous slide

Getting Started

Installation

Important

Please see the Moodle for a guide about how to install Power BI on your laptop!

You should have completed this before class in your own time.

Live Demo

Live Demo

Walkthrough:

Introduce key components of the Power BI interface:

  • The Canvas
  • Fields pane
  • Visualisations pane
  • Filters pane, and
  • Report View.
  • Show you how to import data
  • Show you how to create a basic bar chart
  • Show you how to polish that visualisation

Example

Polishing you visualisation

When you are happy with the plot you’ve chosen for your data it is time to polish the plot.

Steps

  • Update all the plot labels
    • Title, axes labels, tick labels, legend text etc.
    • Use intuitive and intelligent labels
    • You must include units where relevant
    • Ensure the text size is readable!

Polishing you visualisation

Steps

  • Chose an appropriate colour scale

  • Align your visual elements in your plot for clarity

    • e.g. Think about legend position

  • Check the final plot conveys the message you intend

    • Be sure your plot proportions are good
    • Check the range on your axes

  • May like to add additional text or colour to draw the eye to important parts of the plot you want to highlight

Your turn

Your turn

  • Import the same data as me!

  • Try recreating the same plot in Power BI

In Built Visualisations

Close Up: Visualisation Pane

What it does

  • Core interface for building visuals in Power BI.
  • Drag-and-drop fields into Axis, Values, Legend, etc.
  • Default visuals include bar/line charts, tables, maps, pies, treemaps.
  • Supports custom visuals and R/Python integrations.

Barplots

When to use

  • When working with numeric and categorical data

  • Great for comparing differences between categories directly

What to watch out for

  • Not the same as histograms

  • Too many categories create clutter

  • Bars should not be in a random order - Order bars for comparability

  • For readability use horizontal bars

  • Avoid 3D effects

Pie Example - okayish

  • Note the percentages are the percentages related to the top 5 countries as we filtered our data.

Doughnut Example

  • Note the percentages are the percentages related to the top 5 countries as we filtered our data.

Awful Pie Example

Pie and Doughnut Plots

When to use

  • Show proportions of a whole

  • Suitable for small numbers of categories

  • Works best when highlighting a very big or very small segment

  • Doughnut plots are the same as pies (but with holes)

Pie and Doughnut Plots

What to watch out for

  • Humans read angles poorly!

  • Make sure the proportions add up to a whole

  • Labeling wedges directly can be more effective than using a legend

  • Too many slices or similar colours reduce clarity

  • Doughnuts have same drawbacks as pies

Treemap Example

Treemap

When to Use

  • Good for categorical and numeric data

  • Efficient when ther are many categories

  • Also efficient for data with nested structures (hierarchies)

  • Rectangle area encodes value; colour can encode a second variable

Watch out for

  • Limit to ~2–3 hierarchy levels for readability

  • Very small categories become unreadable

  • Interactivity can help when many levels are needed

Your turn

Your turn

  • Try recreating some of these plot types

  • What do you notice about the plots

  • Which ones work best for the data and understanding the key messages?

Concept Check

When you would rather a bar chart, a pie chart or a tree map for visualising data?

Barplots of Multiple Variables

Types of Barplots

About

  • Bar plots can can have a primary and secondary categorical variable.

  • The primary categorical variable can be split into secondary sub-levels. e.g. 1. Country of Birth and 2. Gender

  • Three different types of bar plots

    • Stacked

    • Clustered (grouped), and

    • Percentage stacked

Stacked Barplot Example

Good for comparing the total number of residents.

Stacked Barplots

When to use

  • When comparing totals in the primary category is the most important

  • The composition of secondary levels is also relevant, but less important

Watch out for

  • As the bars are stacked, it is difficult to compare the secondary categories across bars

Clustered (Grouped) Barplot Example

Use to compare differences in the number of residents between census years and for each country.

Clustered (Grouped) Barplots

Note

  • Compare secondary categories within each primary group.

  • Also compare primary categories across secondary groups

Caution

  • Not suitable for comparing totals

  • Can get cluttered with many sub-categories

Percentage Stacked Barplots

Shows the percentage by of residents from each census.

Percentage Stacked Barplots

When to use

  • When interested in proportions within the total

  • Each bar scaled to 100%

  • A variation of stacked bar plots, but better for comparing relative distributions of the secondary groups.

Watch out for

  • No longer comparing the total of the primary category

Small Multiples Example

Small Multiples

About

  • Also known as faceted charts.

  • Used to display multiple visualisations (or “panels”) of the same chart type

  • Each panel represents a subset of data based on a specific category

  • Excellent for spotting patterns across categories

  • Useful when you have a large amount of data - improves data-density

Watch out for

  • Keep axis scales consistent (when appropriate)
  • Maintain consistent colours, labels, and design

Your turn

Your turn

  • Try these different types yourself

Concept Check

  • When you would rather each of these different types? It all has to do with the comparison you want to make!

  • Also important to consider which is the primary and secondary variable when designing bar charts / small multiples

Other Plots

Other Plots

About

Other commonly used plot types are

  • Line charts

  • Area charts

  • Scatter plots

  • Bubble charts

  • Ribbon charts

  • Waterfall charts

  • so on

Map Example

Maps

When to Use

  • Visualise geospatial data (lat/long or regions)

Once the map is drawn, we can:

  • colour each region (choropleth map)

  • add points or bubbles (bubble map)

  • reshape the region (cartogram)

  • show the connection between several regions (connection map)

Watch out for

  • Projection matters for correct areas/distances

What’s missing?

Note

Many common plots that are not in the default visualisation pane:

  • Boxplots

  • Histograms

  • Density plots

  • Contour plots

  • Heatmaps

  • so on

We’ll show you how to create some of these in R

Customisation

About

  • Custom visuals are particularly useful when we need a unique visualisation that is not available in the standard Power BI visual library.

  • We can create custom visuals using R, Python, JavaScript and TypeScript.

  • To create a custom visual in Power BI, we need the Power BI Visuals Tools (pbiviz), which provides a framework for building, testing, and packaging custom visuals.

  • After creating the custom visual, you can import it into your Power BI reports.

Summary

Summary

What we have covered

  • An overview of visualisation in Power BI

  • Covered many of the standard types of visualisations

    • Specifically numeric and categorical

    • Bars, Pies, Dougnuts, Treemaps

    • Stacked, Clustered, and Percent Stacked Bar plots

  • Discussed more about how to choose the right visualisation for the data and the relationships when want to compare

ETX2250/ETF5922