Introduction to data An Andr drew Heiss, PhD Brigham Young - PowerPoint PPT Presentation

Introduction to data An Andr drew Heiss, PhD Brigham Young University visualization with R SLC RUG • July 11, 2018 @andrewheiss

Plan for today Why visualize data? Types of visualizations Aesthetics and design ! Take a sad plot and make it CRAPier !

talks.andrewheiss.com/utah-rug-dataviz/

Why visualize data?

Data alone cannot tell stories or prove theories

Never trust summary statistics alone

Humans are visual creatures @FacesPics

What makes a good visualization?

Characteristics of graphical excellence 1. “... the well-designed presentation of interesting data—a matter of substance, statistics, and design.” 2. Complex ideas communicated with clarity, precision, and efficiency. 3. That which gives the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space. 4. Nearly always multivariate. 5. Requires telling the truth about the data.

��� ����� <latexit sha1_base64="pYen8CoRw76JGC2IK5JDw/YfKA=">ACN3icbVDLSgMxFM34rPVdekmWARXZUYERCKutCNKFgrtKXcSTNtaCYzJHfEMsxfufE3OnGhSJu/QPTB6itBwKHc87l5h4/lsKg6z47U9Mzs3PzuYX84tLymphbf3GRIlmvMIiGelbHwyXQvEKCpT8NtYcQl/yqt896fvVO6NiNQ19mLeCKGtRCAYoJWahYs68ntMTwHh8Fx1qe7rGT2i9UADS4euNbLmTzDLJvXrCEFmWbNQdEvuAHSeCNSJCNcNgtP9VbEkpArZBKMqXlujI0UNAomeZavJ4bHwLrQ5jVLFYTcNLB3RndtkqLBpG2TyEdqL8nUgiN6YW+TYaAHTPu9cX/vFqCwUEjFSpOkCs2XBQkmJE+yXSltCcoexZAkwL+1fKOmALQ1t13pbgjZ8SW52S5b8q72iuXjUR05skm2yA7xyD4pkzNySqEkQfyQt7Iu/PovDofzucwOuWMZjbIHzhf3yfvry4=</latexit> <latexit sha1_base64="pYen8CoRw76JGC2IK5JDw/YfKA=">ACN3icbVDLSgMxFM34rPVdekmWARXZUYERCKutCNKFgrtKXcSTNtaCYzJHfEMsxfufE3OnGhSJu/QPTB6itBwKHc87l5h4/lsKg6z47U9Mzs3PzuYX84tLymphbf3GRIlmvMIiGelbHwyXQvEKCpT8NtYcQl/yqt896fvVO6NiNQ19mLeCKGtRCAYoJWahYs68ntMTwHh8Fx1qe7rGT2i9UADS4euNbLmTzDLJvXrCEFmWbNQdEvuAHSeCNSJCNcNgtP9VbEkpArZBKMqXlujI0UNAomeZavJ4bHwLrQ5jVLFYTcNLB3RndtkqLBpG2TyEdqL8nUgiN6YW+TYaAHTPu9cX/vFqCwUEjFSpOkCs2XBQkmJE+yXSltCcoexZAkwL+1fKOmALQ1t13pbgjZ8SW52S5b8q72iuXjUR05skm2yA7xyD4pkzNySqEkQfyQt7Iu/PovDofzucwOuWMZjbIHzhf3yfvry4=</latexit> <latexit sha1_base64="pYen8CoRw76JGC2IK5JDw/YfKA=">ACN3icbVDLSgMxFM34rPVdekmWARXZUYERCKutCNKFgrtKXcSTNtaCYzJHfEMsxfufE3OnGhSJu/QPTB6itBwKHc87l5h4/lsKg6z47U9Mzs3PzuYX84tLymphbf3GRIlmvMIiGelbHwyXQvEKCpT8NtYcQl/yqt896fvVO6NiNQ19mLeCKGtRCAYoJWahYs68ntMTwHh8Fx1qe7rGT2i9UADS4euNbLmTzDLJvXrCEFmWbNQdEvuAHSeCNSJCNcNgtP9VbEkpArZBKMqXlujI0UNAomeZavJ4bHwLrQ5jVLFYTcNLB3RndtkqLBpG2TyEdqL8nUgiN6YW+TYaAHTPu9cX/vFqCwUEjFSpOkCs2XBQkmJE+yXSltCcoexZAkwL+1fKOmALQ1t13pbgjZ8SW52S5b8q72iuXjUR05skm2yA7xyD4pkzNySqEkQfyQt7Iu/PovDofzucwOuWMZjbIHzhf3yfvry4=</latexit> <latexit sha1_base64="pYen8CoRw76JGC2IK5JDw/YfKA=">ACN3icbVDLSgMxFM34rPVdekmWARXZUYERCKutCNKFgrtKXcSTNtaCYzJHfEMsxfufE3OnGhSJu/QPTB6itBwKHc87l5h4/lsKg6z47U9Mzs3PzuYX84tLymphbf3GRIlmvMIiGelbHwyXQvEKCpT8NtYcQl/yqt896fvVO6NiNQ19mLeCKGtRCAYoJWahYs68ntMTwHh8Fx1qe7rGT2i9UADS4euNbLmTzDLJvXrCEFmWbNQdEvuAHSeCNSJCNcNgtP9VbEkpArZBKMqXlujI0UNAomeZavJ4bHwLrQ5jVLFYTcNLB3RndtkqLBpG2TyEdqL8nUgiN6YW+TYaAHTPu9cX/vFqCwUEjFSpOkCs2XBQkmJE+yXSltCcoexZAkwL+1fKOmALQ1t13pbgjZ8SW52S5b8q72iuXjUR05skm2yA7xyD4pkzNySqEkQfyQt7Iu/PovDofzucwOuWMZjbIHzhf3yfvry4=</latexit> What makes Minard’s graph so great? �������� ����� = ��� ����

We forget this!

Types of visualizations

Exploratory visualizations Academic-ish Quick scatterplots, histograms, other charts to help understand your data Explanatory visualizations Publishable Consumable by the general public; Vox, NYT, Washington Post, FiveThirtyEight, etc .

Exploratory data analysis Find analytical insight in data (even causal inference ! )

Explanatory data analysis Annotate and tell a story

Exploratory Explanatory

Which chart type do I use?

Aesthetics and design

Is he maybe a little right? plot(mtcars$wt, mtcars$mpg)

R can’t do everything There’s still a place for Illustrator, InDesign, et al.

But R can do a lot And it can automate most of the hard work

What constitutes good design?

Four core design principles Co Contrast Re Repetition Al Alignmen ent Pr Proximity

Contrast “If f two it items are not exactly ly th the same, , make th them diffe ifferent. Really different .” .” Don’t be a wimp!

Serif Lorem ipsum dolor sit amet Sans Serif Lo Lorem ipsum um do dolor sit am amet Slab Serif Lorem ipsum dolor sit amet Lo Lorem ip ipsu sum do dolor sit am amet Script Lo Lorem ip ipsum do dolor sit am amet Decorative Light Lorem ipsum dolor sit amet Black Lorem ipsum dolor sit amet

https://color.adobe.com/ http://colorbrewer2.org/ Scientific Colour-Maps viridis https://github.com/thomasp85/scico

Repetition “R “Repeat at some as aspect of f the desig ign throughout the entir ire pie iece.”

Alignment “Every it item should ld have a vis isual l connectio ion wit ith somethin ing els lse on the page.”

Proximity “Group rela lated it items together.”

Contrast Repetition Alignment Proximity

! Take a sad plot and make it CRAPier !

By default, R graphics violate CRAP ggplot(mtcars, aes(x = wt, y = mpg)) + plot(mtcars$wt, mtcars$mpg, geom_point() + main = "Here's a title") labs(title = "Here's a title")

With ggplot’s theme() and other functions, we can make beautiful CRAPy figures automatically with R You can also do this in base R, but I find ggplot’s paradigm more intuitive talks.andrewheiss.com/utah-rug-dataviz/

Moral of the story Graphics are essential for telling stories and gaining insight Design principles (CRAP) make graphics better understandable R + ggplot can follow CRAP and make beautiful, insightful graphics