More effective data visualisation

Visualisation for communication with others

Dr Nicola Rennie

Introduction

Lecturer in Health Data Science within the Centre for Health Informatics, Computing, and Statistics.

Background in statistics, operational research, and data science consultancy.

Co-author of Royal Statistical Society’s Best Practices for Data Visualisation guidance.

CHICAS logo

Why visualise data?

Data visualisation has two main purposes:

Exploratory data analysis and identifying data issues
Communicating insights and results

Examples of sequential, diverging, and qualitative palettes

See nrennie.rbind.io/training-data-visualisation/slides1.html for exploratory data analysis discussion.

Outline

This session will include:

the role of visualisation in communication;
how to design more effective visualisations;
examples of good and bad charts!

Communicating insights with data visualisation

Grab attention

Visualisations stand out. If a reader is short on time or uncertain about whether a document is of interest, an attention-grabbing visualisation may entice them to start reading.

Improve access to information

Textual descriptions can be lengthy and hard to read, and are frequently less precise than a visual depiction showing data points and axes.

Summarise content

Visual displays allow for summarising complex textual content, aiding the reader in memorising key points.

Designing better charts

What are you trying to communicate?

Who are you trying to communicate it to?

Visualisations should be intuitive to the group of people you are communicating with.

Think about your audience

Dissertation presentation?
Dissertation report?
Journal article?
Magazine article

Conference presentation?
Stakeholder meeting?
- NHS data analysts?
- Clinicians?
- Policy makers?

These are different audiences. They might each require a different type of visualisation for the same data.

What makes charts less intuitive?

Type of chart doesn’t match the type of data
Inconsistent colour schemes
Poor choices of axes
Overly complex chart design
Not enough explanation

Some types of charts are less intuitive than others.

Why do people hate pie charts?

Good data visualisation isn’t just based on opinions…

Why do people hate 3D charts?

On the plot on the left, how tall is the bar?

The design process

Questions to answer before you build

What type of chart?
How will you design that chart?
- What orientation is the plot?
- Are there multiple plots?
- Should you use colour?
- What does the title say?

Even after you decide to make a bar chart, for example, there are still a lot of choices to be made about how that bar chart will look.

Start with sketching

Choosing appropriate axes

Ordering discrete variables

Alphabetical order:

Magnitude ordered:

Naturally ordered:

Don’t make it more difficult than it needs to be

Longer labels are best on the y-axis, horizontally.

Plotting multiple values

Avoid spaghetti plots!

Effectively plotting multiple values

Alternatives to spaghetti:

Show a smaller number of lines (e.g. compare a few countries to average)
Use facets (AKA small multiples)
Use colour only to highlight lines

Plotting multiple variables

Effectively plotting multiple variables

Some alternatives:

Separate plots, each with their own axis, and place the plots side-by-side.
Plot different variables on the x- and y- axis.
Rescale the variables, rather than the axis.

Know when lines are appropriate

Suggest an order
Suggest continuity

Sometimes just a number is enough…

… but sometimes it’s not!

Keep it simple, but don’t over simplify.

Discussion 1

03:00

What are the problems with this type of chart?
What could you do instead?

The use of colour

Colours

Why use colours in data visualisation?

Colours should serve a purpose, e.g. discerning groups of data
Colours can highlight or emphasise parts of your data.
Not always the most effective for, e.g. communicating differences between variables.

Different types of colour palettes for different types of data.

Differentiating categories with colour

Before and after scatterplots visualising Ferris wheel data showing only dots, vs, dots and triangles to differentiate groups.

Tip: In a series of charts, assign the same colour to the same variable in each chart.

Are intuitive colours always best?

Example: red and blue used to show hot and cold

Tip: never switch to the opposite meaning!

Are intuitive colours always best?

Example: red and green used to show bad and good

Tip: check the accessibility of colour palettes.

Accessible colours

Check for colourblind friendly plots with colorblindr::cvd_grid(g).

Are intuitive colours always best?

Example: pink and blue used to show women and men

Tip: think about colour associations.

Legends

Order and orientation of the legend matches the order and orientation of the data
Use direct labelling.

The use of text

Why use text?

Just because you’re making a picture doesn’t mean you can’t also use words!

Improve clarity
Explain how to interpret a complex chart
Highlight interesting data points
Summarise a message

Textual elements

Title: A concise description of the main message you want the visualisation to show.

Subtitle and/or Caption: A more specific description of the data used, and how the visualisation should be interpreted. Attribution of the data source. Any caveats about the data or analysis.

Annotations: Highlight specific points, or time periods.

Alt text: An accessible text description of the visualisation, including chart type and the main message.

Choosing a font

Font size: larger fonts are (usually) better
Font colour: ensure sufficient contrast
Font face: highlight text using bold font, avoid italics
Font family: choose a clear font with distinguishable features. Does a 1 look different from a lowercase l, and from an uppercase I?