Visualization

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Debotosh Bhattacharjee
Department of Computer Sc. & Engg
Jadavpur University
 Data Visualization
 Creation and study of visual representation of data

 It involves converting our data source into visual

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

representation

 These might be simply charts, graphs maps etc.

 “The simple graph has brought more information to the data

scientist’s mind than any other device” – John Tukey

FFT & box plot

 What makes an effective Data Visualization?
 It has a clear purpose and message

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

 Complements and enhances texts

 It is easy to interpret

 “A good figure is like a joke, if you have to

explain it, it’s not that good”

 It accurately reflects data

 Creating effective visualization type
 What are you trying to highlight in your data?
 Are you trying to show a variable’s overall pattern of

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

variations or differences among individual values.
 Design of Data Visualization

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Graphics design: Information design:
Communicating idea in an Communicating
engaging and appealing information effectively to
way facilitate comprehension
Prof. Debotosh Bhattacharjee, CSE, Jadavpur University
 Prof. Debotosh Bhattacharjee, CSE, Jadavpur University
 Gestalt Principles are principles or laws of human perception
that describe how humans group similar elements, recognize
patterns and simplify complex images when we perceive
objects.
 Law of similarity
 The law of similarity states that we perceive
elements as a group if they're similar in shape,

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

size, color, or other characteristics.
 Take this image, for example. Image Source.
 When you group items with certain characteristics,
our brains naturally draw links to help us try and
understand the relationship
 Law of similarity

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

 Law of proximity
 The law of proximity is about the physical distance between
objects or things.
 Things or objects that are close together are perceived as a

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

group or unit.
 Things or objects
that are not close
together are not
perceived as
belonging
together.
 Law of enclosure
 The enclosure focuses on how we perceive objects
enclosed in a common area.
 Under this principle, our brains perceive these objects as if

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

they belong to that area.

 As such, the brain

can fill in the
missing pieces of
information to
create a complete
image.
 Law of closure
 The human mind tends to perceive incomplete shapes as
complete forms.
 If an element or shape in a chart or graph is incomplete,

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

our brains will naturally fill in the missing parts to create a
whole, recognizable pattern.
 This can be leveraged to simplify visualizations and
enhance clarity.
 Law of continuity
 The human eye tends to follow the smoothest, most
continuous path when viewing lines or shapes, regardless
of how the lines are actually drawn.

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

 Elements arranged on a line or curve are perceived as more
related than elements not on the line or curve.
 This guides the viewer's eye and organize data, making it
easier to identify patterns and relationships.
 Law of connection
 Objects perceived as connected are grouped together as a
single entity.
 Elements connected by lines, arrows, or other visual cues

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

are interpreted as a single, related group.
 This principle is crucial for effectively showing relationships
between data points and for guiding the viewer's
perception of the data.
 Pre-attentive processing: color
 Color, shape help viewers quickly grasp important data
points without consciously focusing on them.
 Color or shape can be used to highlight differences,

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

emphasize certain data, or convey relationships.
 By strategically using color, shape or both you can make
visualizations more intuitive and effective,
 Pre-attentive processing: color

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Ranking the effectiveness of visual encodings
Magnitude channels Identity channels

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Magnitude channels
encode ordered
attributes and
Identity channels
encode categorical
attributes
 Aesthetics
 Aesthetics. refers to the way a product looks and feels. The
first impression of a product is often visual, making

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

aesthetics very important.

 Whenever we visualize data, we take data values and

convert them in a systematic and logical way into the visual
elements that make up the final graphic.

 All data visualizations map data values into quantifiable

features the make up the aesthetics of the final graphic.
 Aesthetics
 To represent continuous and discrete data
 Use graphical representations- must have a shape,

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

a size, and a color

 Use lines – with

different widths or
dash–dot patterns

 For text, specify font

family, font face,
and font size.
Source: https://clauswilke.com/dataviz/aesthetic-mapping.html#aesthetics-and-types-of-data
 of
A directory

visualizations

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

 Amounts
 The most common plots for visualizing amounts include bar
charts, dot charts, and heatmaps. The amount over time
can be visualized using line charts.

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Heatmap Line chart

Dot plots Bar plots

Amounts

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 22
Amounts

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 23
Visualizing multiple categorical variable

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 24
Visualizing multiple categorical variable
Grouped Bar Plots

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 25
Visualizing multiple categorical variable
Grouped Bar Plots

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 26
Visualizing multiple categorical variable
Stacked Bar Plots

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 27
Visualizing multiple categorical variable
Percent Stacked Bar Plots

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 28
Visualizing multiple categorical variable
Facetted Bar Plots

Prof. Debotosh Bhattacharjee, CSE, Jadavpur University

Dr. D. Bhattacharjee, Department of Computer Sc. & Engg, Jadavpur University 29