The third part of our series on data analytics from digital guru Chris Argent introduces the art of data visualisation.
In one sense, data visualisation is not new to us finance professionals.
Accounting has always involved handling data, from processing transactions, to analysing performance, or complying with regulations and standards
Part of that challenge has been to make the data easy to understand with visualisation.
The big change has been in the sheer amount of information available. And of course the sophistication of tools available to do it.
Organisations generate more data these days. There also public or commercial databases with useful information. As a result, better tools with more power and functionality have been developed. So we need to take a fresh look at data visualisation and consider it as a key tool in our toolkit.
Pictures are power
Data visualisation has been around for longer than you think. And people have been using it to solve big problems and gain a strategic advantage.
Early world maps, such as the Piri Reis map, are a good illustration.
The Piri Reis map was created in 1513 by the cartographer of the same name from military intelligence sources. In those days, the nation that ruled the waves had the potential to rule the world. So a reasonably accurate visualisation of oceans and coastlines had the potential to unlock a huge advantage.
The processes behind the compilation of the map are similar to modern day data visualisation:
- First, define the challenge.
- Second, bring the data together.
Source data for the map was distributed all over the world and not structured in a helpful, easily consumable way. This included navigator knowledge, ships logs, historical diaries, regional map and picture books.
So Piri Reis’s solution was to combine 20-34 sources, remodel them into one consolidated view, understand them better in one database, then create one visualisation using colour, pictures, text and keys.
500 years on, we take much the same approach in data visualisation today.
Design principles for data visualisation
Data visualisation needs to help businesses observe trends and patterns, helping users see those patterns fast, understand and remember the data better, and enable real-time decision making, by speeding up consumption of data and changing course, based on live data feeds.
All of this can be done globally, collaboratively, quickly on a self-service basis, and content can be explored and edited by both accountants and analysts, and business users with their own custom views to help them understand their part of the business decision.
Data visualisation focuses on presenting data in the most meaningful way, for quick delivery of insight to support decision making. It involves making best use of visuals or graphical resources to strengthen human awareness and facilitate understanding of information and insight.
Design for the mind
Great presentation requires a little understanding of the way the human mind works and how we perceive and react to information.
- People find multiple attributes hard to process.
- Our attention capacity is limited.
- Cluttering our field of vision is distracting.
- Patterns can be observed if presented effectively.
- People’s attention is drawn to contrast.
- We dislike ambiguity.
Example: Hollywood films
This popular visualisation was created to explain the performance of the biggest Hollywood films. It was well received, but in terms of business decision making it helps to highlight some of our own limitations with consuming data and the need to get it right.
How accurately and quickly can you answer the below performance-related questions?
- What’s the ratio of profitable films or loss-making films?
- Which genre has the most flops?
- Is there any relationship between the Critic’s rating and profit recovery?
Example: probability matrix
A probability impact matrix like the one below is a very common way of visualising project risk. It demonstrates good use of contrast. For example, red draws attention to the high risk projects that need to be reviewed. But it’s also a bit bland. Also, it’s a little ambiguous because it doesn’t tell a story or give a precise message about what to focus on.
How accurately and quickly can you answer the below performance related questions?
- Which project needs our most attention?
- What the definition of a high-risk project?
- Are any “moderate” impact projects defined as high-risk projects?
Example: life expectancy around the world
This is a more successful example of data visualisation. The problem it addresses is clear and the illustration also tells a story about record-keeping over time and by location. You can probably detect this and answer the following questions without difficulty.
From these graphics, how easily could you answer the following questions?
- Which regions have the best and worst life expectancy?
- Which countries have been keeping records the longest?
- How has average life expectancy changed?
For more information on Data Analytics, we recommend these articles.
How to develop your skills
There is a growing range of software to help with visualisation projects. But even with these, it is vital to broaden your own understanding as much as you can.
There are several ways to do this.
- Be aware of all the different types of visualisations available to you.
- Buy a few books that “make data beautiful”.
- See how infographics and graphic designers make data more interesting to their readers.
As with the Piri Reis map, think about the problem and the purpose of the visualisation and go on a journey with the message.
- Think of the user and allow them to find the message on their own terms, with their preferences.
- Consider the overall challenge, audience needs, and actual content.
- How should you structure the display for speed and simplicity?
- What can you do to ensure messages can be consumed fast?
- Finally, what will visualisation be?
Make a conscious decision on the type of visualisation that you want to use, and how easy it will be for users to understand the message in the data. Think of the pro’s and con’s of using one visualisation compared to another, one colour scheme to another. It’s worth taking time to get the above right, as the visual may be part of a report or dashboard for a very long time. A few minutes here, my save a lot of questions later on.
The next level – data dashboards
Individual visualisations can be combined in to one big visual to highlight the most important information and KPIs needed by each type of business user. The aim is to create a single view that can be understood at a glance. Enter the Dashboard.
Dashboards are hard to get right. As you can see, just one visual needs careful consideration so a collection of visuals with different formats and messages needs even more. The main challenges are to display all the required metrics and messages on one screen, clearly, so it’s easy to understand and consume fast.
It is possible if worked on collaboratively. Consider the overall design keeping the user in mind, ask them want they want regularly and aim to eliminate clutter and distraction, group data into logical sections consider the messages being communicated, highlight what’s most important (the top left of a screen is the most prominent place for this) and show meaningful comparisons, not just last year or the budget, think of decision lead times and use that time horizon.
Take Data Visualisation seriously. As you can see people like to see information in a particular way, and the above is just the tip of the data/mind iceberg. It’s time to get serious as data will only get bigger and bigger, and getting data visualisation right can be the difference between making a good business decision and a bad business decision, winning a profitable deal or losing a deal, retaining your business seat at the table or losing them.
KPIs – Key Performance Indicators a quantifiable measure used to evaluate the success of a business and organisation in meeting strategic and performance objectives.
Data Attribute – Is the characteristic of data that sets it apart from other data, such as location, length, or type. This is important to understand when visualising lots of data from one model.
Data Model – Is how we organise different data attributes and simplify and standardise how they relate to one another in the real world entities.
Common Data Visualisation Tools – Sisense, Looker, Domo, Tableau, Power BI, Qlik, MicroStrategy.
This article was written by Chris Argent and was first published in AAT Comment. It forms the third article in a series of five, access them all here:
Data Analytics Series
- Data analytics – 1 – here’s what you need to know to get started
- Data analytics – 2 – your role in an analytics-focused finance team
- Data analytics – 3 – visualisation techniques to bring your data to life
- Data analytics – 4 – how data science and machine learning fit in
- Data analytics – 5 – how to communicate your insights