Data Visualization Quiz 3
Eliminate Clutter
Include only what adds to the audience's understanding, remove everything else
aes()
aesthetic mapping, map data to elements of the chart
ggplot
create any chart, need to add various elements on at a time for display
qplot
create basic charts, can send aesthetic mappings in as parameters
data-ink ratio
data-ink/total ink used to print graphic
avoid diagonal components including lines and text
e.g. people read text rotated at 45 degrees ~50% slower than normal text
Gestalt Principle: connection
elements tend to be grouped together if they are connected by another element
Gestalt Principle: similarity
elements tend to be integrated into groups if they are similar to each other
Gestalt Principle: enclosure
elements tend to be integrated into groups if they are within an enclosure
Gestalt Principle: closure/continuity
elements tend to be integrated into groups if they are within an enclosure. your brain fills in the rest of the shape even though parts are removed
Gestalt Principle: proximity
elements tend to be perceived as aggregated into groups if they are near each other. strongest principles
Clutter
elements that take up space/use ink but do not increase understanding
we "see" with a combination of our
eye(s)
often default behavior in data visualization tools create...
more contrasting elements than are needed
without visual cues...
most readers will read in a z-shaped pattern from left to right, then down a line, and so on.
whitespace means
no non-data-ink
Data-ink
non-erasable ink used for the presentation of data (i.e. if any of the data-ink is removed, the graphic would lose content)
intensity (saturation)
ordered data with a few values, ordinal data, broadly supported, will generally require a legend to define the order, be careful of visual habits e.g. temperature ranges often represented by a red to blue scale
line width
ordered data with a few values, ordinal data, supported by some data viz tools (your instructor thinks this one is silly)
orientation
ordered data with a few values, quantitative data, ordinal data, not commonly supported
Size
ordered data with many useful values, quantitative data, ordinal data, broadly supported data viz tools, sometimes size is used when shape is a better encoding
line length
ordered data with many values, quantitative data, ordinal data, this preattentive attribute explains why bar charts are very effective for their use, broadly supported in data viz tools
position
ordered data with many values, quantitative value, ordinal data, categorical data, relationship data, broadly supported, note: listed twice once as a way to reduce cognitive load and again as a way to draw attention
Create clean horizontal and vertical lines in the design
otherwise readers will read left to right, then down and so on
ggplot2
package created by Hadley Wickam to increase the functionality of plotting in R
Gestalten
plural form of Gestalt
applicable gestalts with whitespace
proximity, enclosure
5 gestalt principles
proximity, similarity, enclosure, closure/continuity, connection
the data-ink ratio represents...
the proportion of the graphics ink devoted to the non-redundant display of data/information
consider the following and use whitespace liberally:
tufte's data-ink ratio
directing the audience
"enhancing the understanding" by providing direction
removing clutter
"not hurting" your audience
Chartjunk
"useless, non-informative, or information-obscuring elements of quantitative information displays"
light comes into...
the lens and focuses on the retina
Edward Tufte quote #2
Above all else, show the data
Chartjunk quote credit
Edward Tufte
theme
allows the user to control the overall look and feel
ggplot is flexible...
and extensible way to generate graphics
the eye-brain system has...
approx 1000x the resolution of a standard power point slide
Throughput is approx. 10 Mbps(each)
brain
retina is made from
brain cells (seeing and thinking are one at this point)
"Editor"
chooses, forgets, summarizes, focus, etc.
facets
generates the same plot for different slices of the data. will cover in lab 4
examples of geom
geom_point (scatterplot), geom_bar (bar chart), geom_line (line chart), geom_hline (add a horizontal line)
geom
geometrical objects added to layers of the chart.
Every element added....
increases cognitive load
2 main functions to generate plots
qplot, ggplot
gestalt principles
rules of the organization of perceptual scenes.they aim to formulate the regularities according to which the perceptual input is organized into unitary forms, also referred to as (sub)wholes, groups, groupings, or Gestalten (the plural form of Gestalt)
guideline: use preattentive attributes
sparingly; too many reduces focus for the audience
process steps:
start with everything "untreated", "treat" aspects of the data viz one at a time with an effective preattentive attribute, stop before you add to many "treatments"
contrasts directs...
the audience
scales
this set of functions allows the user to control the axes
3-8 seconds
time to get reader's attention
ggplot2 allows the user
to programmatically control nearly every element of the chart
hue (color)
unordered data with a few values (5 is getting to be too many), categorical data, broadly supported, don't over-do it many data viz tools wil run out of different colors and then wrap the colors so two different elements can be represented by the same color (confusing!), be aware that a portion of your audience may be color blind and won't be able to distinguish between certain color combinations.
shape
unordered data with a medium number of values, categorical data, broadly supported, make sure shapes are mapped to the same data across different visuals (otherwise, confusion follows)
added marks
unordered data with a medium number of values, categorical data, mostly represented by call-outs in charts
enclosure
unordered data with many values, categorical data, relationship data, most often represented by call-out boxes overlaying some portion of the chart
results from this step travel...
via the optic nerve to the brain. Optic nerves
Gestalt
your brain filling in the gaps