DataVis 2015 – Augsburg

Note: on the Mac, I used Chrome; Safari didn’t behave properly with this presentation.

(Not) Random Musings

• What is meant by “interesting”?
  • Objectively interesting? Is there such a thing?
  • Personally interesting?
  • Interesting as determined by metrics of internet activity?
  • Does the label “interesting” depend on assumptions which might be incorrect?
• What is meant by “right” in the world of data analysis?
  • Hadley: “The first thing is [rarely] the right thing.”
  • Here, we have a refreshing diversity of perspectives and opinions.

The plan

• 3 short examples
• One longer example for you to explore and critique
• Conclusion

Example 0: My First Collaboration with the Media (2007-ish, Wall Street Journal)

1. WSJ techies ran a cron job querying opentable.com hourly for a few weeks.
2. They pulled me in. I found that “interesting”.
3. I handled the import, tidying, and exploration.
4. They politely said thank you, and then promptly modified my plots in Adobe Illustrator.

Example 1: Graphical eye candy

Bryan Lewis suggested that term when he gave me a first demonstration of htmlwidgets. Ease of technology? Fascinating. Cool? Sure! But… is the data visualisation itself “interesting”?

suppressWarnings(suppressMessages(library(maptools)))
library("threejs")
load("flights.RData")
data(wrld_simpl)

earth <- tempfile(fileext=".jpg")
jpeg(earth, width=2048, height=1024, quality=100,
     bg="#000025", antialias="default")
par(mar = c(0,0,0,0), pin = c(4,2), pty = "m",  xaxs = "i",
    xaxt = "n", xpd = FALSE, yaxs = "i", bty = "n", yaxt = "n")
plot(wrld_simpl, col="black", bg="#000025", border="cyan", ann=FALSE,
     setParUsrBB=TRUE, lwd=3)
dev.off()

## quartz_off_screen 
##                 2

globejs(earth, arcs=flights[sample(nrow(flights), 5000), ])

Example 2: How quickly the world changes

http://www.stat.yale.edu/~jay/world.html

How would this best be done circa 2015? Circa 2012, I used an R script to produce the necessary country overlays as individual .png files as well as the Javascript. What is “interesting” here?

Example 3: Environmental Performance Index 2014

I provide handouts just in case we don’t all have access to the web. However, web exploration is preferable to merely thumbing through the handout, so please do try to connect:

http://epi.yale.edu

1. Front page
2. Second page: the “Data Explorer”
3. Three “country profiles” for Switzerland, Germany, and China
4. Three “issue rankings” for Slovenia, the Solomon Islands, and the USA, in the area of Biodiversity and Habitat (an issue area with four so-called indicators)

Visualization technologies don’t mislead people – people mislead people

I’ll begin with some examples, including a project of my own that engaged a professional web developer. Such examples can lead to further exploration and discussion of topics ranging from academic perspectives and pedagogy to real-world data visualisation. Some of these extend beyond an isolated conversation of graphical interactivity. For example, the dangers of random explorations and misleading interpretations are not unique to interactive graphics. When problems occur, we need to take the time to stop and ask, “Who, really, is at fault? What can we do? What should we be doing?” A well-known quote from the US National Rifle Association comes to mind: “Guns don’t kill people – people kill people.” We focus on technology – graphical innovations in this case – and ignore pedagogy at our own peril.