Wednesday, December 2, 2009

Data Visualization Workshop presented by Monica Bulger & Aaron Sobel

During the workshop, Monica and Aaron shared considerations and techniques for effective data visualization and showed examples of figures that could use improvement from last year's GP posters. Bren's resident visual design experts, Professor Jim Frew and Darren Hardy, also shared insights into effective visual design. The information imparted during the workshop will be useful for GP defenses, public presentations, reports, and posters, as well as other assignments and publications.


1. Who is your audience? What are their needs and level of understanding?
- Clients, advisors, students, members of the public

2. What is your purpose?
- Advocate, inform, convince

3. Figures should be able to stand alone; do not assume the readers will read all your text, they may just look at the figures.
ex) In an otherwise excellent GP poster, a group included a pie chart with relative proportions of survey results. However, they did not include who was surveyed, how many people were surveyed, what question(s) they were asked, and the context of the survey.
ex) Place a map in your description of a study area and include an inset with the larger state/regional context.
ex) When one figure was on the screen, we had to ask, "What do the black dots mean?" Avoid this problem by including everything in your legend.

4. Figures should convey not just the data itself, but the meaning of the data.

5. Decide what story you want to tell with your data. If you don't tell the audience your own data story, they'll make up their own!
ex) A figure of Earth at night shows which areas are lit at night. It tells the story of population, our preferred locations, electricity use, development, etc.

6. Find the best way to represent your data, and prioritize truth and clarity.
ex) Figure size and structure must be readable at a reasonable distance.
ex) When choosing the significance of colors, think of what stands out (light colors) and if colorblind people be able to understand your figure. In a map series from a GP poster, the most significant parts on the maps were black, which did not stand out against a blue background.

7. Sketch your visual design before you even touch Powerpoint or Keynote. Often your tool will dictate your design when clarity should be your first priority.

8. "Slides should be processed in 3 seconds or less. It's impossible for people to process your slides and your words simultaneously." ~Nancy Duarte, author of slide:ology
[Takeaway message: Monica says the same is true for data graphs, charts, etc. -- they should stand alone, be compelling, and be quickly understood by your intended audience.]

Group Exercise:
We then conducted an exercise in which we suggested how to improve a figure on "The Pathway to Self-Funding" for the spiny lobster fishery. Here is a list of specific points that emerged:
- What is the significance of the colors?
- Is the funnel/bowl shape significant? It looks like a beer funnel.
- Things don't flow up, so the visual is illogical.
- The title doesn't match the image.
- Highlight the necessary components for the self-funding policy.
- The box labels don't mean anything to the reader.
- The direction and curves in the arrows add unintentional meaning--if it's a linear process, don't complicate it with curved arrows.
- Since the "next steps" take up more space than the steps the group completed, the figure minimizes the amount of work the group put into their project.
The exercise took 10 minutes, most of which was spent deciphering, which is too long to understand a figure.

Data Visualization Techniques:
- A cartogram is an effective and interesting way to show frequencies by location
- 3D graphs overlayed on maps
- Stacked graphs can show waxing/waning frequencies
- State/regional map insets provide context for the study location
- Draw the readers' attention by highlighting important cells in tables
- Decision trees and timelines make it easy to follow a course of events

Other Resources:
- slide:ology available in the Communications Center and the library server
- Professor Jim Frew recommends Tufte's texts on visual design and the following data visualization websites: and

Coming soon: discussion of tools for visual design....

--Sara Solis

No comments:

Post a Comment