These slides are adapted from the Coursera JHU Data Science Specialization course material available through a Creative Commons license and the Swirl course repository available through a GNU GPL v3.

Thanks to the original authors for making their material available for reuse and remixing!

These slides are made available under a Creative Commons BY-NC-SA license, similar to the material from which they were derived.

• Principle 1: Show comparisons

• Evidence for a hypothesis is always relative to another competing hypothesis.

• Always ask "Compared to What?"

• Principle 1: Show comparisons

• Evidence for a hypothesis is always relative to another competing hypothesis.

• Always ask "Compared to What?"

• Principle 2: Show causality, mechanism, explanation, systematic structure

• What is your causal framework for thinking about a question?

• Principle 1: Show comparisons

• Evidence for a hypothesis is always relative to another competing hypothesis.

• Always ask "Compared to What?"

• Principle 2: Show causality, mechanism, explanation, systematic structure

• What is your causal framework for thinking about a question?

• Principle 3: Show multivariate data
• Multivariate = more than 2 variables
• The real world is multivariate

• Need to "escape flatland"

• Principle 4: Integration of evidence

• Completely integrate words, numbers, images, diagrams

• Data graphics should make use of many modes of data presentation

• Don't let the tool drive the analysis

• Principle 4: Integration of evidence

• Completely integrate words, numbers, images, diagrams

• Data graphics should make use of many modes of data presentation

• Don't let the tool drive the analysis

• Principle 5: Describe and document the evidence with appropriate labels, scales, sources, etc.

• A data graphic should tell a complete story that is credible

• Principle 4: Integration of evidence

• Completely integrate words, numbers, images, diagrams

• Data graphics should make use of many modes of data presentation

• Don't let the tool drive the analysis

• Principle 5: Describe and document the evidence with appropriate labels, scales, sources, etc.

• A data graphic should tell a complete story that is credible

• Principle 6: Content is king

• Analytical presentations ultimately stand or fall depending on the quality, relevance, and integrity of their content

• Principle 1: Show comparisons
• Principle 2: Show causality, mechanism, explanation
• Principle 3: Show multivariate data
• Principle 4: Integrate multiple modes of evidence
• Principle 5: Describe and document the evidence
• Principle 6: Content is king

• To understand data properties
• To find patterns in data
• To suggest modeling strategies
• To "debug" analyses
• To communicate results

• To understand data properties
• To find patterns in data
• To suggest modeling strategies
• To "debug" analyses
• To communicate results

• They are made quickly
• A large number are made
• The goal is for personal understanding
• Axes/legends are generally cleaned up (later)
• Color/size are primarily used for information

• The U.S. Environmental Protection Agency (EPA) sets national ambient air quality standards for outdoor air pollution

• U.S. National Ambient Air Quality Standards

• For fine particle pollution (PM2.5), the "annual mean, averaged over 3 years" cannot exceed \(12~\mu g/m^3\).

• Data on daily PM2.5 are available from the U.S. EPA web site

• EPA Air Quality System

• Question: Are there any counties in the U.S. that exceed that national standard for fine particle pollution?

One dimension

• Five-number summary
• Boxplots
• Histograms
• Density plot
• Barplot

summary(pollution$pm25)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##   3.383   8.549  10.050   9.836  11.360  18.440

boxplot(pollution$pm25, col = "blue")

hist(pollution$pm25, col = "green")

hist(pollution$pm25, col = "green")
rug(pollution$pm25)

hist(pollution$pm25, col = "green", breaks = 100)
rug(pollution$pm25)

boxplot(pollution$pm25, col = "blue")
abline(h = 12)

hist(pollution$pm25, col = "green")
abline(v = 12, lwd = 2)
abline(v = median(pollution$pm25), col = "magenta", lwd = 4)

barplot(table(pollution$region), col = "wheat", 
        main = "Number of Counties in Each Region")

Two dimensions

• Multiple/overlayed 1-D plots (Lattice/ggplot2)
• Scatterplots
• Smooth scatterplots

\(> 2\) dimensions

• Overlayed/multiple 2-D plots; coplots
• Use color, size, shape to add dimensions
• Spinning plots
• Actual 3-D plots (not that useful)

boxplot(pm25 ~ region, data = pollution, col = "red")

par(mfrow = c(2, 1), mar = c(4, 4, 2, 1))
hist(subset(pollution, region == "east")$pm25, col = "green")
hist(subset(pollution, region == "west")$pm25, col = "green")

with(pollution, plot(latitude, pm25))
abline(h = 12, lwd = 2, lty = 2)

with(pollution, plot(latitude, pm25, col = region))
abline(h = 12, lwd = 2, lty = 2)

par(mfrow = c(1, 2), mar = c(5, 4, 2, 1))
with(subset(pollution, region == "west"), plot(latitude, pm25, main = "West"))
with(subset(pollution, region == "east"), plot(latitude, pm25, main = "East"))

• Exploratory plots are "quick and dirty"

• Let you summarize the data (usually graphically) and highlight any broad features

• Explore basic questions and hypotheses (and perhaps rule them out)

• Suggest modeling strategies for the "next step"

• Edward Tufte (2006). Beautiful Evidence, Graphics Press LLC. www.edwardtufte.com