Introduction to Data Visualisation

Introduction to Data Visualisation

Data visualisation is the graphical representation of data and information. According to Kirk (2016), it is both a technical practice and a communicative act — transforming raw numbers and figures into visual forms that can be understood, interpreted, and acted upon by audiences. Crucially, visualisation is never neutral: every design decision shapes what the audience sees, what they understand, and what they are likely to do with that understanding.

Understanding Data: The Foundation

Before a dataset can be visualised, it must be understood. Kirk (2016) identifies several dimensions of understanding data that shape the visualisation process:

• What type of data is it — categorical, ordinal, quantitative, temporal, spatial?
• What relationships exist within the data — comparisons, rankings, proportions, distributions, trends?
• What is the story the data tells, and what story do you want to communicate?
• What are the limitations, gaps, or biases in the data?

These questions determine which visualisation types and design approaches are most appropriate. A chart that mismatches its form to its data type will confuse or mislead rather than inform.

Ways of Seeing Data

Gray, Bounegru, Milan, and Ciuccarelli (2016) propose an influential three-stage critical framework for analysing how data visualisations are made and received. They call it Ways of Seeing Data, and it moves through three stages:

1. World to Data — This stage examines what information and data are being represented, and how that data came to exist. Key questions include: What are the sources of this data? How was it generated? What methods and standards are inscribed in the data infrastructure? How has the data been transformed or prepared? Which data sources have been combined? How does the data selectively prioritise certain things over others? This stage reminds us that data is never a simple mirror of the world — it is always a partial and constructed representation.
2. Data to Image — This stage examines how data is mediated into graphical form. Key questions include: What graphical techniques and technologies were used? What are their affordances — what do they make visible, and what do they obscure? What design decisions were taken, and what are their consequences for how the data is understood? This is where visual grammar, chart selection, colour, layout, and typography all come into play.
3. Image to Eye — This stage examines how audiences encounter and interpret the final visualisation. Key questions include: What visual cultures and practices are reflected in this design? Who are the intended publics of this visualisation? How is it circulated, cited, and shared? How do different viewers decode it in different ways?

Symbolic Meaning of Charts

Charts and graphs carry symbolic meaning that extends beyond their data content. Aiello (2020) explains this through the concept of semiotic regimes: sign-making in data visualisation is regulated by social practices, authority, expertise, and conformity within particular contexts. The meaning of visual forms is not fixed or universal — it is produced through culturally specific systems of convention.

Aiello (2020) frames this through the idea of a visual grammar: a system of rules for combining visual elements to create meaning. As Machin (2007) puts it, a visual grammar provides a 'lexicon of elements that can be chosen to create meaning in combinations' and 'a finite system of rules' for their combination. Cultural and social histories shape the meaning of specific visual elements — like colour, shape, and layout — and tracing these histories helps us understand how meaning is made and changed over time.

Visual Grammars: Common Chart Types

Different chart types carry different visual grammars and are suited to different communicative purposes:

• Bar charts — compare single discrete values across categories; convey relative magnitude clearly and efficiently
• Pie charts — communicate segments of a whole; effective for showing proportions, though often misused when there are too many segments
• Line charts — show change over time; particularly effective for continuous data and trends
• Timelines — organise information chronologically; useful for historical narratives and sequences of events
• Choropleth maps — colour-coded maps for geographical data; use spatial metaphors to communicate regional variation
• Multiple maps — show location-specific data across different time points or categories for comparative purposes
• Scatter graphs — display relationships between two variables; identify correlations, clusters, and outliers
• Bold headings and icons — provide emphasis, hierarchy, and navigational clarity within complex visualisations

Visual Communication and Pictograms

Pictograms and icon-based visualisations translate data into recognisable symbolic forms. They are particularly effective for communicating with broad, non-specialist audiences because they leverage existing visual knowledge and cultural familiarity. Pictograms were used in early data communication and remain a common feature of contemporary infographics, particularly in public health, environmental reporting, and civic communication.

Data Stories and Narrative Structure

Effective data communication involves more than accurate charts — it requires narrative. Data stories organise information into a meaningful arc that guides the audience through what the data shows, why it matters, and what it implies.

Segal and Heer (2010) divide the design space of data storytelling into three elements:

1. Genre — the overall form of the visualisation narrative (e.g. magazine style, annotated chart, partitioned poster, flow chart, comic strip, slide show, video)
2. Visual narrative tactics — specific techniques for guiding attention and creating flow (e.g. highlights, annotations, progressive disclosure)
3. Narrative structure — the overall organisation of the story, from linear (beginning-middle-end) to non-linear (exploratory, user-directed)

Choosing the right narrative structure depends on the purpose and the audience. A persuasive visualisation for a general audience may benefit from a clear linear narrative with a definitive conclusion. An exploratory tool for expert analysts may allow non-linear navigation. Understanding this design space enables more deliberate and effective data storytelling.