bettervis

Making Better Graphs

Prerequisites

• Basic R programming experience
• No familiarity with ggplot2 required
• Experience working with data

Learning outcomes

• How to think about creating good plots
• Good vs bad charts
• Data-ink ratio
• Clear message communication
• Hierarchy of plot design (concepts of proximity)
• Fundamentals of ggplot2:
  • Understanding the grammar of graphics
  • Working with different geoms and aesthetics
  • Common pitfalls and how to avoid them
  • Polishing plots for publication
  • Brief introduction to useful common extensions (e.g., patchwork)

Learning Outcomes

By the end of this course, participants will be able to:

1. Understand the grammar of graphics and how ggplot2 implements it
2. Map variables to aesthetics (x, y, colour, fill, shape, size) appropriately
3. Choose appropriate geometries for different data types (points, lines, bars, histograms)
4. Use faceting strategically to create small multiples with facet_wrap() and facet_grid()
5. Apply the proximity principle when designing multi-variable plots
6. Recognise and create tidy data suitable for ggplot2
7. Reshape data using pivot_longer() and pivot_wider()
8. Create effective bar plots with proper sorting and positioning
9. Polish plots with labels, themes, colours, and legends
10. Handle overplotting using jitter, transparency, and summaries
11. Save publication-quality figures with ggsave()

Course website

https://better-vis.njtierney.com

This course goes through not just how to create graphics, but how to communicate your message in your graphic.

Schedule

Session 1: Introduction & Scatterplots

• Philosophy and approach to understanding ggplot2
• Learn the anatomy of a ggplot using the template approach
• Map variables to aesthetics (x, y, colour, shape, size)
• Understand the difference between mapping aesthetics inside aes() versus setting them outside
• Practice with variables-aesthetics.R and colour.R

Session 2: Faceting & Small Multiples

• Learn to create small multiples with facet_wrap() and facet_grid()
• Understand when to use single vs multiple variable faceting
• Apply the proximity principle: put the comparisons you want readers to make close together
• Explore different orderings of colour vs facets to change interpretations
• Practice recreating plots with different faceting strategies using facet.R

Session 3: Tidy Data & Line Graphs

• Introduction to tidy data principles and why they matter for ggplot2
• Learn to identify if data is tidy using the template: variables, observations, and values
• Reshape data with pivot_longer() and pivot_wider()
• Create line graphs with geom_line() using gapminder data
• Understand the group aesthetic for multiple series
• Practice with tidy-data.R and gapminder-line.R

Session 4: Bar Plots & Histograms

• Create bar plots with geom_bar() and understand when to use geom_col()
• Learn the importance of sorting with fct_infreq() and fct_rev()
• Understand fill vs colour for 2D shapes
• Use position options (stack, dodge, fill) for grouped bar plots
• Create histograms with geom_histogram() and choose appropriate bins or binwidth
• Handle overlapping histograms with alpha and position = "identity"
• Practice with barplots.R and histogram.R

Session 5: Polishing Your Plots

• Learn the polishing workflow: save plots to objects, then add labels, improve colours, apply themes, and adjust legends
• Add comprehensive labels with labs()
• Use colourblind-safe colour palettes: scale_colour_brewer(), scale_colour_viridis_d(), and colorspace palettes
• Apply built-in themes like theme_minimal() and extension themes from ggthemes and hrbrthemes
• Control legend position with theme(legend.position)
• Save publication-quality plots with ggsave()
• Practice with polish.R

Session 6: Advanced Topics & Practice

• Handle overplotting with geom_jitter(), controlling width and using alpha for transparency
• Understand boxplot anatomy and when to use summaries
• Learn why summaries lose detail (DataSaurus revisited)
• Explore alternatives like half-and-half plots with gghalves and raincloud plots with ggrain
• Advanced tricks: adding background data layers to faceted plots, and plotting many variables at once with pivot_longer() and faceting
• Open practice time and Q&A session

This site is open source. Improve this page.