Principles of data visualisation 2020
Download as PPTX, PDF1 like486 views
This document provides guidance on principles of data visualization. It discusses why we visualize data, such as to communicate findings and inspire action. The visualization process involves getting and cleaning data, setting goals, and choosing visual types based on the data and audience. Effective use of color, narrative, and networks are also covered. The document emphasizes knowing the audience to select the right visual type and story to engage them. Overall it provides a helpful overview of best practices for data visualization design and communication.
Principles of data visualisation 2020
- 1. Marié Roux (Manager: Research Impact Services) Photo by ThisisEngineering PRINCIPLES OF DATA VISUALISATION
- 2. CONTENT • Why do we visualise data? • Know your audience • Find the story • Choose visualisation type • Effective use of colour • Use of narrative • Visualising networks • References • Useful links
- 4. Data Visualisation process Getting data Know the data Setting goals Visualise or not? What to visualise Cleaning data Visualising Designing Exporting/ embedding
- 5. Why do we visualise data? • Primary reason to visualise – Communicate a point or a finding revealed in your analysis • To add legitimacy or credibility, tell stories with numbers • Better understand the data you gathered • Data more persuasive when shown in graphs or visualisations • To inspire others into action • Effective data visualisation can increase the impact of your research and your engagement efforts Only a picture can carry such a volume of data in such a small space Edward Rolf Tufte
- 6. Why do we visualise data? (Cont.) Two broad goals according to Angela Zoss: 1. Visualisation for analysis: • Explore and analyse data relationships • Try many views and combinations to find meaningful stories 2. Visualisation for communication: • Select a particular view of the data to share • Construct the visualisation with a goal in mind and taking the audience into account, for example when a decision needs to be made
- 7. 3 Types of Data Visualisation Data visualisation is an umbrella term, usually covering both information and scientific visualisation. This is a general way of talking about anything that converts data sources into a visual representation (like charts, graphs, maps, sometimes even just tables). Scientific visualisation: generally, the visualisation of scientific data that have close ties to real-world objects with spatial properties. The different scientific fields often have very specific conventions for doing their own types of visualisations. Information visualisation: also a broad term, covering most statistical charts and graphs but also other visual/spatial metaphors that can be used to represent data sets that don't have inherent spatial components. Infographic: a specific sort of genre of visualisations. Infographics have become popular on the web as a way of combining various statistics and visualisations with a narrative.
- 8. Visual literacy Visual literacy skills equip a learner to understand and analyse the contextual, cultural, ethical, aesthetic, intellectual, and technical components involved in the production and use of visual materials. A visually literate individual is both a critical consumer of visual media and a competent contributor to a body of shared knowledge and culture. In an interdisciplinary, higher education environment, a visually literate individual is able to: • Determine the nature and extent of the visual materials needed • Find and access needed images and visual media effectively and efficiently • Interpret and analyse the meanings of images and visual media • Evaluate images and their sources • Use images and visual media effectively • Design and create meaningful images and visual media • Understand many of the ethical, legal, social, and economic issues surrounding the creation and use of images and visual media, and access and use visual materials ethically ACRL Visual Literacy Competency Standards for Higher Education
- 11. “When you tell the right story to the right audience, and are able to identify data points that the specific audience can relate to and encourage them to start a conversation, you increase your story’s share-ability and give it the chance of going viral” Jessica Dubow, Amanda Makulec.
- 12. Know your audience Understand your audience before designing your visualisation The first and most important consideration is your audience. Their preferences will guide every other decision about your visualisation—the dissemination mode, the graph type, the formatting, and more. You might be designing charts for policymakers, funders, the general public, or your own organisation’s leaders, among many others. What type of decisions do your viewers make? What information do they already have available? What additional information can your charts provide? Do they have time (and interest) to explore an interactive website, or should you design a one-page handout that can be understood at a glance? Example: A chart designed for local government leaders wouldn’t be appropriate for a group of programme implementers, and vice versa.
- 13. Know your audience (2) Remember that the key is to keep your audience engaged If you are sharing results in client meetings, staff retreats, conferences, or webinars, try breaking up your charts into several slides so the chart appears to be animated. This is called a storyboarding technique. Draw Attention to key charts with handouts Curated handouts might be useful to draw attention to specific highlights of your data. Tweeting your results If you are planning to tweet a chart or two, be sure to adjust your charts to fit a 2:1 aspect ratio. Otherwise your visualisation will get chopped in half because when you are scrolling through your Twitter feed, the images automatically display about twice as wide as they are tall. Consider these questions: How do you engage your team when creating and presenting data for your team or organisation? What types of communications modes are you currently using to share your visualisations?
- 14. Know your audience (3) Ask two questions: - Who is the data visualisation intended for? - What does the audience know about the topic? Also consider the audience’s level of - Literacy: You can use symbols, illustrations, animations and other universally understood graphics - Numerical literacy: Even educated audiences are not always comfortable with data and stats. Do they understand ratios, complex formulas or statistics. Or do they need simplified data? - Education/Level of technical expertise: Simplify content and define terms for less technical audiences but provide more detail for those with expertise - Job function: What is the purpose of your data visualisation? Academics will want to know how data fits into existing literature. Funders will want to see results compared to money spent. An executive will want high-level results to guide decisions. A programme manager may only be interested in data relevant to their specific topic/area.
- 15. Find the story
- 16. Find the story Data Visualisation as storytelling - Think of your data visualisation message as a thesis statement which needs a summary in a few concise sentences - The ability to create a compelling, visual argument will be greater if you begin with a clear and focused message - The type of story you tell affects the platform you will use. - Infographics might be more useful for persuading the audience of your point of view, where dashboards leave the interpretation to the audience. - Animations may be effective for college students, but not amongst older adults. Examples: Maps: John Snow’s map of 1854 cholera outbreak, identify point of origin, influence policy makers to improve water sanitation. Graphs: Florence Nightingale’s diagram of causes of death in the Crimean War, more soldiers died from preventable illnesses than battle wounds, allocated more resources and training to health workers Infographics: For the general audience without background knowledge, data is simplified by visuals and is not numerical intimidating, engaging with broad audience. It follows a well-defined story, key messages highlighted, clear purpose.
- 22. Visualisation types 1D/Linear • lists of data items, organized by a single feature (e.g., alphabetical order) (not commonly visualised) 2D/Planar (incl. Geospatial) • Choropleth • Cartogram • Dot distribution map • Proportional symbol map • Contour/isopleth/isarithmic map • Dasymetric map • Self-organizing map Sources: https://guides.library.duke.edu/datavis/vis_types https://uark.libguides.com/dataviz
- 23. Visualisation types (2) 3D/Volumetric • 3D computer models • Surface and volume rendering • Computer simulations Temporal • Timeline • Time series • Connected scatter plot • Gantt chart • Stream graph/ThemeRiver • Arc diagram • Polar area/rose/circumplex chart • Sankey diagram • Alluvial diagram Source: Wikipedia. Steamgraph
- 24. Visualisation types (3) Multidimensional Source: https://images.app.goo.gl/fSiB9ZPdoQke4ymb8 • Pie chart • Histogram • Wordle, tag cloud • Tree map • Scatter plot • Bubble chart • Line chart • Step chart • Unordered bubble chart/bubble cloud • Bar chart • Radial bar chart • Area chart/stacked graph • Heat map • Parallel coordinates Radar/spider chart • Box and whisker plot • Mosaic display • Waterfall chart
- 25. Visualisation types (4) Tree/Hierarchical • General tree visualization • Dendrogram • Radial tree • Hyperbolic tree • Tree map • Wedge stack graph (radial hierarchy)/sunburst • Icicle/partition chart Network • Matrix • Node-link diagram • Dependency graph/circular hierarchy • Hive plot • Alluvial diagram See detailed lists of different types of visualisations: https://datavizproject.com/ https://datavizcatalogue.com/index.html Source: https://datavizproject.com/data-type/hyperbolic-tree/
- 27. Use of colour Use of colour is a subjective topic, but there are some basic principles to think about when choosing the colours of your visualisations. One colour theory concept: the HSL model. HSL breaks colour down into three separate channels: hue, saturation and luminance. Hue – is what most people think of as colour – red, blue, yellow, green, purple, etc. Each colour is plotted on a scale from 0° to 359° to form a colour wheel. Saturation – is another word for a colour’s intensity. The scale measures how different the colour looks from neutral gray, which has 0% saturation. Colours with high saturation look brighter and more vivid. In the example above, you can see how saturation increases towards the bottom right corner of the triangle. Luminance – describes the spectrum of a hue from dark, based on the amount of black added. In our example above, luminance increases towards the bottom right left of the triangle. Source: https://cambridge-intelligence.com/choosing-colors-for-your-data- visualization/
- 28. Use of colour (2) See how choosing different colours can make a difference. Source: https://cambridge-intelligence.com/choosing-colors-for-your-data- visualization/
- 29. Use of colour (3) Play this game: Color Your goal is to match colours.
- 30. Use of colour (4) How to choose colours Step 1: Decide what the colours will represent Decide which aspects of your data you want to represent with colour. Step 2: Understand your data scale The ColorBrewer tool defines three types of scale: • Sequential – when data values go from low to high • Divergent – when data has data points at both ends of the scale, with an important pivot in the middle. • Qualitative – when the data does not have an order of magnitude.
- 31. Use of colour (5) How to choose colours Step 3: Decide how many hues you need Based on the scale you chose in step 2, you can decide how many hues you need in the palette: • Sequential data usually requires one hue, using luminance or saturation to define scale. • Divergent data requires two hues, decreasing in saturation or luminance towards a neutral (usually white, black or gray). • Qualitative data requires as many hues as values, but remember the limitations of the human brain. Try to not use more than seven or eight colours, otherwise the brain cannot recall what each one represents. Step 4: Look for obvious options Before getting too creative, take a look at your data to see if there’s an obvious set of colours. Your application or corporate style guide might be a good starting point. See example on next page.
- 32. Use of colour (6) How to choose colours Example: Weather temperatures. Here blue and red are understood without explanation
- 33. Use of colour (7) How to choose colours Step 5: Create your palette Use one of the many web resources. ColorBrewer is one of the best for picking schemes for sequential, diverging and qualitative data. Or if you have a starting point in mind, Adobe Color creates palettes from a single colour. There are several groups of colours that work well together. You can identify them by their relative positions on the colour wheel: • Monochromatic – shades of a single hue (sequential data). • Analogous colours – colours that sit beside each other on the colour wheel (varied alternative for sequential data). • Complementary colours – from opposite sides of the colour wheel (diverging data). • Triadic colours – 3 colours equally spaced around the wheel (good starting point for a qualitative palette).
- 34. Use of narrative
- 35. Use of narrative Ways to visualize qualitative data Qualitative data consist of non-numeric information that doesn't involve measurements or quantities. Coding Qualitative Data Qualitative data analysis often use the method of marking a certain section of data with a particular code when it falls into a specific theme or category. Tools like Atlas Ti and Dedoose can assist with coding and visualisation. Text analysis Analysis of the structure and linguistic features of a collection of text in order to detect measurable patterns. Like coding, text analysis allows us to visualise qualitative information in a quantitative way. Simple text analysis can be performed with a user-friendly tool like Voyant, but more complex analysis may require a programming language such as Python. Source: https://guides.lib.unc.edu/DataViz/qualitative
- 36. Use of narrative Ways to visualize qualitative data Illustrative Diagrammes If you want to produce a data visualisation but you don't have any kind of dataset to work from, you most likely need to create an illustrative diagramme. Diagrammes that involve basic shapes such as flowcharts and mind maps can easily be made with Microsoft Powerpoint. Detailed or complex illustrations are usually created with professional design software like Adobe Illustrator. Visualisation methods Word Clouds, Heat Maps and Prezi can be used to visualise coded text and text analysis.
- 38. Visualising networks Network analysis examines relationships between different entities, such as collaborations between researchers, interactions between genes, or communications between people in a company. It can be used for a wide range of purposes from simply studying the structure of a community to solving complex math and engineering problems. Network visualisation is the visual component to network analysis. There are a wide variety of network visualisation types to choose from depending on what type of data you have or what types of relationships you want to show. Source: https://guides.lib.unc.edu/DataViz/networks
- 39. https://vimeo.com/18477762 Hans Rosling says there’s nothing boring about stats, and then goes on to prove it. A one-hour long documentary produced by Wingspan Productions and broadcast by BBC, 2010.
- 40. References • A Reader on Data Visualization. MSIS 2629 Spring 2019. Santa Clara University. https://mschermann.github.io/data_viz_reader/ • Dubow, Jessica & Amanda Makulec. 2014. Identifying your audience and finding your data story. JSI Center for Health Information. • Evergreen, Stephanie. 2017. Effective data visualization: the right chart for the right data. Los Angeles: Sage Publications. • Kelleher, C & Wagener, T. 2011. Ten guidelines for effective data visualization in scientific publications. Environmental Modelling & Software. Volume 26, Issue 6, June, Pages 822- 827. • Olshannikova, E., Ometov, A., Koucheryavy, Y., & Olsson, T. 2016. Visualizing Big Data. In Big Data Technologies and Applications (pp. 101-131). Springer International Publishing. • Quinn, Laura. 2017. Creating great data visualizations with low cost tools. July. Idealware. • WILS. Principles of data visualization. https://goo.gl/eMRHOU • WinWire Technologies. 2017. Principles for creating effective data visualization. 22 Aug 2017. • Zoss, Angela. Data Visualization: About Data Visualization. Duke University. Library guide. • Zoss, Angela. 2012. Introduction to data visualization. (Slideshare). • University of North Carolina. Guide to Data Visualization: Qualitative. Library guide.
- 41. Useful links • https://www.techchange.org/2015/05/21/audience-matters-in-data-visualization/ • https://playfairdata.com/vital-question-1-who-is-the-audience/ • https://infogram.com/blog/how-data-visualization-impacts-engagement/ • https://flowingdata.com/category/tutorials/ • https://d3js.org/ • https://datavizproject.com/ • https://flourish.studio/2018/09/28/choosing-the-right-visualisation/ • https://extremepresentation.com/wp-content/uploads/choosing-a-good-chart-09-1.pdf • http://labs.juiceanalytics.com/chartchooser/index.html • http://www.ala.org/acrl/standards/visualliteracy • https://library.duke.edu/data/data-visualization • https://blog.datawrapper.de/colorguide/ • https://cambridge-intelligence.com/choosing-colors-for-your-data-visualization/ • https://www.sciencedirect.com/topics/social-sciences/network-analysis
- 42. Questions? Please e-mail any comments, questions to me [email protected] Register for the next online workshop on data visualisation Monday 30 March, 11:00-13:00 Overview of software for effective data analysis and data visualisation Details about the online workshop will be distributed to participants before the time
Editor's Notes
- #4: Data Visualisation is more than just plotting graphs and pretty charts. It’s important to deliver a message effectively. Edward Rolf Tufte is an American statistician and professor of political science, statistics, and computer science at Yale University. “Only a picture can carry such a volume of data in such a small space”