Introduction: This chapter introduces data visualisation, its definition, core principles, and relationship with other disciplines.
1.1 The Components of Understanding
Definition of Data Visualisation: "The representation and presentation of data to facilitate understanding."
Importance of the Definition: The components of this definition are crucial and will be referenced throughout the book.
Data: The raw material for visualisation.
Without data, there is no visualisation.
Data includes names, amounts, groups, statistical values, dates, comments, and locations.
Data is typically in tabular form with rows (records) and columns (variables).
Tables allow precise reading of individual data points but make it difficult to establish comparative relationships.
Data Representation: The act of visually portraying data through charts, marks, and attributes.
Marks: Points, lines, and areas used in charts.
Attributes: Appearance properties of marks, such as size, color, and position.
Chart Anatomy: Marks, attributes, axes, and gridlines that form a chart.
Data Presentation: All other visible design decisions beyond representation, including interactivity, annotation, color usage, and composition.
Stages of Understanding
Goal: Data visualisation aims to facilitate understanding.
Three Stages of Understanding: Perceiving, interpreting, and comprehending.
Each stage is dependent on the previous one.
The visualiser has influence but not full control over these stages.
Perceiving
The ability to read a chart and decode data representations (shapes, sizes, colors) into perceived values.
Interpreting
Converting perceived values into meaningful insights based on pre-existing knowledge.
The visualiser can assist with captions, headlines, and colors to clarify meaning.
Comprehending
Reasoning the consequences of perceiving and interpreting to form a personal reflection.
Involves relating new information to prior knowledge and considering its relevance and impact.
Demonstration of Understanding
Example: Analyzing Lionel Messi's career statistics using a clustered bar chart.
Process: Commences with perceiving the chart, followed by interpreting the data, and finally comprehending the analysis.
Detailed Breakdown of the Process
Perceiving: Identifying chart type, axes, labels, and color legend. Scanning and observing physicalproperties, noting prominent variations in size, shape, color, and position. Identifying big, small, and medium values (magnitude judgements).
Interpreting: Converting perceived readings into meaning by orienting assessment against existing knowledge. Understanding data in relation to the subject.
Comprehending: Concluding reasoning that translates into what the analysis means personally, inferring from the display of data, and relating/responding to the insights drawn through interpretation.
1.2 The Importance of Conviction
Design Process: Following a data visualisation design process helps make good design decisions.
Guiding Principles: Rely on good design principles for nuanced situations.
Influences: Early beliefs about data visualisation design are often shaped by prominent authors in the field.
Examples: Edward Tufte, Stephen Few, David McCandless, Alberto Cairo, and Tamara Munzner.
Dieter Rams' Principles
Dieter Rams: German industrial and product designer known for his work with Braun.
10 Principles of Good Design:
Good design is innovative.
Good design makes a product useful.
Good design is aesthetic.
Good design makes a product understandable.
Good design is unobtrusive.
Good design is honest.
Good design is long lasting.
Good design is thorough down to the last detail.
Good design is environmentally friendly.
Good design is as little design as possible.
Principles for Data Visualisation Design
Three high-level principles translated from Rams' principles:
Good data visualisation is trustworthy.
Good data visualisation is accessible.
Good data visualisation is elegant.
Principle 1: Good Data Visualisation is Trustworthy
Trust: The fundamental integrity, accuracy, and legitimacy of any data visualisation.
Trust vs. Truth: Truth is an obligation, while trust is earned. There is rarely a singular view of the truth in data visualisation.
Securing Trust: Eliminate any sense that your version of the truth can be legitimately disputed.
A visualisation can be truthful but not viewed as trustworthy, and vice versa.
Examples: Comparing graphics from the UK Office for National Statistics (ONS) and the Daily Mail.
ONS graphic uses low-key colors and includes important explanatory features, earning more trust.
Daily Mail's graphic feels gimmicky and lacks detailed data source information.
Principle 2: Good Data Visualisation is Accessible
Accessibility: Removing design-related obstacles to facilitate understanding.
Reward vs. Effort: Minimizing friction between the act of understanding (effort) and achieving understanding (reward).
Human-Centered Design: Demonstrating empathy for your audiences and putting them at the heart of your decision-making.
Factors Your Audiences Influence
Subject-matter appeal: This is a fundamental junction at the beginning of the consumption experience.
Dynamic of need: Is engagement voluntary or necessary?
Subject-matter knowledge: What might your audiences know and not know about this subject?
What do they need to know?: Often, the most common frustration expressed by viewers is that the visualisation ‘didn’t show them what they were most interested in’.
Unfamiliar representation: A key challenge lies with situations when the deployment of an uncommon chart may be an entirely reasonable and appropriate choice – indeed perhaps even the ‘simplest’ chart that could have been used – but it is likely to be unfamiliar to the intended viewers.
Time: At the point of consuming a visualisation is the viewer in a pressured situation with a lot at stake?
Format: What format will your viewers need to consume your work?
Personal tastes: Preferences towards certain colors, visual elements and interaction features will often influence (enabling or inhibiting) a viewer’s engagement
Attitude and emotion: The prospect of working on even the most intriguing and well-designed project sometimes feels too much
Factors Influenced by Visualisers
Solution is Useless: Failure to focus on relevant content, an oversimplified complex subject, or work that requires too much time to make sense of, when immediate understanding and rapid insights were needed.
Solution is Obtrusive: Visually inaccessible, misjudged format, or too many functions.
Solution is not understandable: Complex subject or analysis, uses a complex chart type, and/or is absent of annotations
Principle 3: Good Data Visualisation is Elegant
Elegant Design: Achieving a visual quality that attracts the audience and sustains that sentiment throughout the experience.
Elegant Design qualities: Stylish, dignified, effortless and graceful
Achieving Elegance: In serving the principles of trustworthy and accessible design, elegance may have arrived as a byproduct
Key components
Eliminate the arbitrary: A dedicated visualiser should be prepared to agonise over the smallest details and want to resolve even the smallest pixel-width inaccuracies and avoid complacency.
Thoroughness: Avoid going through the motions and don’t get complacent. Often you will find yourself working alone on a data visualisation project and will therefore need to demonstrate the discipline and competence to challenge yourself
Style: The decisions around colour selection, typography and composition are all matters that influence your style so don't be too concerned about cosmetics.
Decoration should be additive, not negative: Visual embellishments are, in moderation and when discernibly deployed, effective devices for securing visual appeal and preserving communicated value.
Not about minimalism: Elegant design achieves a certain invisibility: as a viewer you should not see design, you should see content
1.3 Distinctions and Glossary
Consistency: Preserve clarity for readers.
Distinctions
Data vis: abbreviated term for data visualisation.
Information visualisation: Common use of data visualisation and vice versa, also used as the term to define work that is primarily concerned with visualising abstract data structures such as trees or graphs (networks) as well as other qualitative data
infographics: Traditionally created for print consumption, in newspapers or magazines. Contain charts (visualisation elements) but may also include illustrations, photo-imagery, diagrams and text
Visual analytics: Serve the role of operational decision support systems or provide instruments of business intelligence/analytical reasoning and exploration of data facilitated by interactive tools
Data art: Artists goal is not driven by facilitating the kind of understanding that a data visualisation would offer (more about pursuing a form of self-expression or aesthetic exhibition using data as the paint and algorithms as the brush).
Information design: Much broader application concerned with the design of many different forms of visual communication.
Data science: A field, hard to define, which is easier to consider through the ingredients of the role of data scientists (gathering, handling and analysing of data).
Data journalism: Also known as data-driven journalism (DDJ), this concerns the increasingly recognised importance of having numerical, data and computer skills in the journalism field.
Scientific visualisation: The label given exploratory data analysis (drawing out the scientific methods for analysing and reasoning about data) and others relate to it the use of visualisation for conceiving highly complex and multivariate datasets specifically concerning matters with a scientific bent.
Roles and Terminology
Project:
Visualiser:
Viewer:
Audience:
Consuming:
Creating:
Data Terminology
Data is:
Raw data:
Dataset:
Tabulation:
Variables:
Series:
Data source:
Big Data:
Visualisation chart Types
Chart type:
Graphs, charts, plots, diagrams and maps:
Graphic:
Storytelling:
Format:
Function:
Axes:
Scale:
Legend:
Outliers:
Correlation:
Summary Defining Data Visualisation
Chapters Summary: Chapter covered the principles and components of data visualization.