Data Visualization & Storytelling

Session Overview

  • Publications:
    • What makes an effective publication?
    • How do we tell a story?
  • Visualizations:
    • Best Practices
    • Figure Formats
    • Storytelling

Publications

  • Publications:
    • What makes an effective publication?
    • How do we tell a story?


Writing Science

Scientific Storytelling

  • Memorable
  • Efficient
  • Effective

Story

  • Content
  • Structure
  • Language

Content

Memorable

  • Simple
  • Unexpected
  • Concrete
  • Credible
  • Emotional
  • Stories

Visualizations

  • Best Practices
  • Figure Formats
  • Storytelling

Visualization Best Practices


Show what you mean to show

nothing more, nothing less.


Examples


Leverage Code and Editing Appropriately

Capture the Statistics

  1. Central Tendency
  2. Variation
  3. Extreme Values

Match Dimensional Rendering


  • If you are in 2 dimensions, your data should be displayed in 2 dimensions.
  • More dimensions?
    • Dimension reduction
    • Parallel Coordinates Plot

The Basics

  • Proportionality (default in many plots)
  • Orient your viewer
    • Axes, legends, etc.
  • KISS
    • Just because you can, should you?
    • One figure one message…
    • Do you really need three axes?
  • Accessibility
    • Color blindness
    • Lighting?
    • Text Readers?

Session Overview

  • Best Practices
  • Figure Formats
  • Storytelling

Figure Formats

Exporting your Figures


  • Write out your figures directly from your script to a figures directory in the desired vector format.

  • Set proportion in your code.

  • Use vector editor for subsequent edits (Illustrator, Keynote), not raster editors (PPT, word)

  • Submit/use high quality vector graphics in your presentations, papers, and posters if possible.

Session Overview

  • Best Practices
  • Figure Formats
  • Storytelling

Storytelling

Visualization Storytelling

“Making sure a picture is worth 1,000 words.

  • Structure your visualizations in a story arc.

  • If a reader only looked at your figures, would they understand your project?

  • Use storytelling arc with figures

Visualization Context


We are going to focus on 2D static figures primarily for presentations and publications


Other areas:

  • Interactivity
  • Multi-dimensionality
  • Exploratory
  • Web

Python

HoloViz/Datashader

  • Supports all major python graphing libraries

  • Native support for Dask

  • Integrates with GeoPandas and Cartopy

  • Scales…

Examples…

Earth System Data Science in the Cloud