FusionCharts

FusionCharts

White Paper

Principles of Data Visualization -

What We See in a Visual
Principles of Data Visualization - What We See in a Visual

Table of Contents
Why We Visualize Data ............................................................................................................................................................ 3

The Importance of Visualizations in Business ................................................................................................... ............ 3

Two Goals in Data Visualization ........................................................................................................................................... 4

Explain Data to Solve Specific Problems .................................................................................................................... 4
Explore Large Data Sets for Better Understanding ................................................................................................. 5

Eyesight and Memory Working in Parallel ....................................................................................................................... 6

The Role of Memory in Vision ........................................................................................................... ............................. 7

Preattentive Attributes Used by Our Working Memory .............................................................................................. 8

Forming Analytical Patterns Out of Preattentive Attributes ...................................................................................... 9

Using the Gestalt Principles to Bring Out Patterns in Visualizations ..................................................................... 11

An Example from Recorded Future ........................................................................................................ .......................... 12

Gestalt Principles ........................................................................................................... ......................................................... 13

Preattentive Attributes ........................................................................................................... .............................................. 14

Conclusion ........................................................................................................... ..................................................................... 14

About FusionCharts ........................................................................................................... .................................................... 15

FusionCharts 2
Principles of Data Visualization - What We See in a Visual

Why We Visualize Data

We visualize information to meet a very basic need - to tell a story. It’s one of the most primitive forms
of communication known to man, having its origins in cave drawings dated as early as 30,000 B.C., even
before written communication, which emerged in 3,000 B.C. Vision is the single most important faculty
we use to communicate information.

With the passing of time, we’ve found new ways to visualize information. Today, we’re familiar with the
basic chart types like the line chart, bar chart, and pie chart. However, we rarely stop to think about
why they’re more effective than bland tables, text, and numbers. Further, we can’t easily spot instances
when they’re done wrong, and can be improved. This white paper aims to give you a foundational
understanding of how we process visual information. You’ll be able to use this knowledge when de-
signing your next dashboard, and have an informed opinion on how to communicate even more clearly
and powerfully using visualizations. Even if you don’t design visualizations yourself, these principles
will equip you with skills to better analyze visualizations you come across in the newspapers, or in your
daily work reports. Let’s start with an understanding of why data visualizations are so important in a
business context today.

The Importance of Visualizations in Business

A visual can communicate more information than a table in a much smaller space. This trait of visuals
makes them more effective than tables for presenting data. For example, notice the table below, and
try to spot the month with the highest sales.

Month Jan Feb Mar Apr May Jun

Sales 45 56 36 58 75 62

This data when visualized gives you the same information in a second or two.

FusionCharts
FusionCharts www.fusioncharts.com
| www.fusioncharts.com 35
Principles of Data Visualization - What We See in a Visual

Edward Tufte, a data visualization expert, says ‘Graphical excellence is that which gives to the viewer
the greatest number of ideas in the shortest time with the least ink in the smallest space.’ This trait of
visualizations is what makes them vital to businesses.

Two Goals in Data Visualization

In a business environment, visualizations can have two broad goals, which sometimes overlap.

Explanatory
Exploratory

Explain Data to Solve Specific Problems

Visuals that are meant to direct the viewer along a defined path are explanatory in nature. The bulk of
business dashboards that we come across in day-to-day scenarios fall in this category.

Here, the viewer starts with a question in mind. For example, ‘which country has the highest value?’.
She then views the visualization, and finds an answer to the question.

FusionCharts 4
Principles of Data Visualization - What We See in a Visual

This type of visualization is used in many scenarios for the following tasks:

Answer a question. E.g., How much sales did we have last quarter?
Support a decision. E.g., We need to stock more football jerseys as they were sold out on
most days last week
Communicate information. E.g., Revenue is on track for this quarter
Increase efficiency. E.g., ‘Technical specifications’ is the most viewed section in the product
page. It should be given more visibility.

Explore Large Data Sets for Better Understanding

Exploratory visuals offer the viewer many dimensions to a data set, or compares multiple data sets with
each other. They invite the viewer to explore the visual, ask questions along the way, and find answers to
those questions. For example, below is a visualization of the State of the Union address of recent Presi-
dents.[1]

Here, the viewer starts by familiarizing herself with the visualization, then identifies an area of interest.
For example, which President spoke more about jobs? She then explores the ‘Jobs’ section of the visuali-
zation, and finds her answer. She could then move on to exploring other areas of the visualization.

FusionCharts 5
Principles of Data Visualization - What We See in a Visual

Exploratory analysis can be cyclical without a specific end point. Viewers can find many insights from a
single visualization, and interact with it to gain understanding rather than make a specific decision. This
type of visualization can accomplish the following tasks:

Pose new questions

Explore and discover

Though they’re not as popular as the previous category, exploratory visualizations have gained promi-
nence in recent years with the rise of big data. The high volume of data, and varied data sets that have
become common today lend themselves easily to exploratory analysis.

Now that we’re aware of what makes visualizations intrinsic to business, let’s dive into the mechanics of
how we process visual information. We’ll understand the role of memory in perceiving visual information,
and how to apply that understanding as we work with visualizations.

Eyesight and Memory Working in Parallel

If you’re familiar with big data terminology you would have come across the term ‘massively parallel
processing’ (MPP), as most commonly seen in MapReduce technology that powers Hadoop. It breaks data
down into small units, and processes each of the units in parallel. A similar process has always been used
by us to process information with our vision as well. When we look at a visual, our eyes and our brain
work in parallel to take in new information, and break it into small chunks. Then both the eyes and the
brain process the chunks in parallel to find meaning. Let’s look at an example to better understand this.

Let’s say we walk into a supermarket to buy oranges. Our eyes first scan the layout of the supermarket.
At the same time, our brain processes the various sections of the layout, and instructs the eyes to zone
in on the fruits section. It does this by sending signals about how fruits look from memory. The eyes then
break the entire scanned area into parts, and scans each part to spot the fruits section. The same process
is repeated till we zero in on the oranges in the fruits section. This process of visualizing information is
performed by the eyes and memory working in parallel.

FusionCharts 6
Principles of Data Visualization - What We See in a Visual

The Role of Memory in Vision

While that’s an overview of how we process visual information, let’s discuss the vital role of memory in
our vision. There are two types of memory that come into play when we process visual information.

Long-term memory
Working memory

Long-term Memory

This type of memory is what makes us always expect the units to be marked off on the X and Y axes
of a chart, or the date range selector at the top of a dashboard. This type of memory is formed by past
interactions and experiences. It needs to be considered when designing the layout of a dashboard or
visualization. There should be good reason to go contrary to long-term memory when deciding the basic
structure of a dashboard or visualization. A discussion of long-term memory is out of the scope of this
white paper. Instead, we’ll focus on the other type - working memory.

Working Memory

When we look at a line chart, or notice a number in a dashboard, we use our working memory to store
just the information we need at the moment. This type of memory breaks the entire visual into small
chunks of information, in a process appropriately called ‘chunking.’ Surprisingly, for all the complexity of
our brain, our working memory can hold only about 3 chunks of information at any given time. When
perceiving a complex visual, we’re constantly replacing the 3 available slots in our working memory.
This is why when designing a visual or a dashboard, one of our goals is to limit the number of prominent
chunks of information that we want the user to notice, and pack as much information as possible into
each chunk. While we want to avoid information overload, a sound understanding of this concept can
help us direct the attention of the viewer in a natural way.

Next, Let’s discuss working memory in depth. Just as any language has alphabets, which are combined to
form words, the language of data visualization consists of certain basic building blocks called preatten-
tive attributes.

FusionCharts 7
Principles of Data Visualization - What We See in a Visual

Preattentive Attributes Used by Our Working

Memory
Colin Ware, Director of the Data Visualization Research Lab at the University of New Hampshire, terms
the basic building blocks of the visualization process as ‘Preattentive attributes.’ These attributes are what
immediately catch our eye when we look at a visualization. They can be perceived in less than 10 millisec-
onds, even before we make a conscious effort to notice them. Here’s a list of the preattentive attributes:

Form
Orientation Line Length Line Width Size

Shape Curvature Added Marks Enclosure

Color Spatial Position

Intensity Hue 2-D Position

These attributes come into play when we analyze any visualization. Of this list, only Position, and Length
can used to perceive quantitative data with precision. The other attributes are useful for perceiving other
types of data such as categorical, or relational data. For example, both the pie chart and the bar chart be-
low show the same data. But you can’t easily tell from the pie chart which is the biggest pie. That’s more
clearly visible in the bar chart as it calls on the preattentive attribute of length.

FusionCharts 8
Principles of Data Visualization - What We See in a Visual

Considering preattentive attributes can help when deciding which chart type to use for our data. While
preattentive attributes are what we immediately identify in a visual, they aren’t the only things we notice.
We go on to form patterns out of the preattentive attributes.

Forming Analytical Patterns Out of Preatten-

tive Attributes
If preattentive attributes are the alphabets of visual language, analytical patterns are the words we form
using them. We immediately identify the preattentive attributes in a visualization. We then combine the
preattentive attributes to seek out analytical patterns in the visual. Here are the basic analytical patterns
that we identify when looking at a visual:

FusionCharts 9
Principles of Data Visualization - What We See in a Visual

These patterns are an intrinsic part of our vision, and are frequently used when we analyze and describe
a chart. However, as a word of caution, we are hard-wired to look for patterns in any visual information
we notice. Sometimes we do this even when there isn’t an apparent connection or pattern in the visual.

We’ve seen how preattentive attributes and patterns enable us to process visual information. However,
when crafting visualizations we often want to highlight certain patterns over others. In these cases, Ge-
stalt’s Principles come in handy.

FusionCharts 10
Principles of Data Visualization - What We See in a Visual

Using the Gestalt Principles to Bring Out

Patterns in Visualizations
Gestalt principles describe how our mind organizes individual elements into groups. We can use these
principles to highlight patterns that are important, and downplay other patterns. The image below
illustrates the principles of Gestalt which are relevant to visualization.

Proximity Similarity Enclosure Symmetry

Closure Continuity Connection Figure & ground

Here’s what we notice from each of the illustrations:

Proximity: We see three rows of dots instead of four columns of dots because they are closer
horizontally than vertically.
Similarity: We see similar looking objects as part of the same group.
Enclosure: We group the first four and and last four dots as two rows instead of eight dots.
Closure: We automatically close the square and circle instead of seeing three disconnected
paths.
Continuity: We see one continuous path instead of three arbitrary ones.
Connection: We group the connected dots as belonging to the same group.
Symmetry: We see three pairs of symmetrical brackets rather than six individual brackets.
Figure & ground: We either notice the two faces, or the vase. Whichever we notice becomes
the figure, and the other the ground

FusionCharts 11
Principles of Data Visualization - What We See in a Visual

These principles can help us perform many tasks such as reduce the noise from charts, choose the ideal
aspect ratio, and show relationships between elements more clearly. The Gestalt principles are a power-
ful tool to have when crafting visualizations.

Now that we have a good understanding of how visualizations work, let’s apply these concepts to a real-
world example.

An Example from Recorded Future

We’ll consider the following visualization from Recorded Future, a web intelligence company.[2] This
visualization compares the mentions of Apple, Google, and Microsoft across the web. The data here is
forward-looking, and considers only mentions related to the next five years.

This visualization features two chart types - An area chart, which is grayed out in the background, and a
bubble chart, which is color-coded in the foreground. Let’s analyze this simple visualization, and identify
which elements from this white paper it uses.

FusionCharts 12
Principles of Data Visualization - What We See in a Visual

Gestalt Principles
Figure & ground
The first thing you notice when looking at this visualization is that the bubbles stand out against the
backdrop of the area charts. This shows the Gestalt principle of figure & ground.

Proximity
Zoning in on the bubbles shows 3 distinct groups of bubbles. We can identify this easily because of
how close the bubbles are to each other.

Similarity
Further, we notice that the bubbles are of three colors - green, purple, and blue. This similarity brings
out the grouping even more clearly.

Analytical Patterns
Going up, going down, remaining flat
This pattern is most visible in the area chart. We notice that the overall trend for Apple is upward, while
Google’s Microsoft’s stays flat.

Wide, narrow
There are many peaks along the area chart, but one of them for Microsoft is particularly noticeable in
Jan 2015. This peak is narrower than the other peaks.

Tightly, loosely distributed

We notice the loose distribution of bubbles for all three companies around 2018 and 2019.

Normal, abnormal
Zoning in on the bubbles for Microsoft, we notice an outlier for Microsoft in mid-2015. The other bub-
bles fall within a more normal range.

FusionCharts 13
Principles of Data Visualization - What We See in a Visual

Preattentive Attributes
Spatial position
We use the preattentive attribute of position to track the rise and fall of the area chart. Similarly, we
notice the abnormal bubble in Microsoft’s chart because of it’s higher position compared to the other
bubbles.

Size / Area
The bubbles vary in size. Their size corresponds to the number of web mentions for a particular topic.
This makes it easy to spot the important mentions, and explore them in detail.

Hue / Color
As mentioned earlier, the color of the bubbles makes it easy to classify them into three groups. This em-
ploys the preattentive attribute of Hue.

Intensity
Finally, the low intensity of the area chart places it in the background, giving priority to the bubbles.

Being aware of the preattentive attributes, analytical patterns, and Gestalt Principles can make a visual
come alive to us. Considering we analyze visual information on a daily basis, and how important it is to
business, it pays to know how visualizations work.

Conclusion
In this white paper, we dissected the process of visual perception. First, we saw that all visualizations
have a goal - explanatory, or exploratory. Second, preattentive attributes are the basic building blocks of
a visualization, and are identified by us almost immediately. Third, preattentive attributes lead us to spot
patterns in a visual. Finally, when crafting visualizations we could use the Gestalt Principles to prioritize
important patterns, and downplay the noisy ones.

In conclusion, it’s obvious that we are naturally hard-wired to visualize information in a certain way. Un-
derstanding those basic principles of data visualization will help us craft outstanding visualizations, and
tell compelling stories, much to the delight our colleagues, and end users.

FusionCharts 14
Principles of Data Visualization - What We See in a Visual

About FusionCharts
FusionCharts Suite XT is the industry’s leading enterprise-grade charting component with delightful
JavaScript (HTML5) charts that work across devices and browsers (including IE 6, 7 and 8). Using it, you
can create your first chart in under 15 minutes and then add advanced reporting capabilities like drill-
down and zoom in a couple of hours after that. It comes with extensive docs, demos and personalized
tech support that makes sure implementation is a breeze for you.

22,000 customers and 500,000 developers in 120 countries including organizations like NASA, Micro-
soft, Cisco, GE, AT&T and World Bank use FusionCharts Suite XT to go from data to delight in minutes.

Learn more about how you can add delight to your products at www.fusioncharts.com

FusionCharts 15
Principles of Data Visualization - What We See in a Visual

Reference:
1. The Washington Post. History Through the President’s Words [1]
2. Recorded Future. Apple, Google, and Microsoft during the next five years [2]

Further Reading:
1. Tufte, Edward. (2001). The Visual Display of Quantitative Information, 2nd Edition. Graphics
2. Press.Ware, Colin. (2008). Visual Thinking. Morgan Kaufmann.
3. Ware, Colin. (2012). Information Visualization, 3rd Edition. Morgan Kaufmann.
4. Few, Stephen. (2009). Now You See It: Simple Visualization Techniques for Quantitative Analysis. Analytics Press.

FusionCharts 16