Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.2

Unit Syllabus

• Data Science and Big Data Fundamentals

• Introduction to Data
• Understanding Data & Exploration
• Data Pre-Processing
• Data Processing
• Model Selection and Evaluation
Introduction to Data Science and Big Data
Topic 9: Lecture 1
• Data Visualization
Data Visualization • Business Intelligence and Tools
• Data Science Ethical and Privacy Issues
• Unit Summary

© NCC Education Limited

Data Visualisation Topic 9 - 9.3 Data Visualisation Topic 9 - 9.4

Scope and Coverage Learning Outcomes

This topic will cover:
• Data Visualization
By the end of this topic students will be able to:
• Build big data analysis solutions with analysis methods
• Data Visualization Process and visualisation tools.
• Advanced data visualization • Understand the need of businesses and present data.
and processing tools • Select appropriate visualisation tools and visualise
aspects of the data in each dataset.
• Systematically interpret and evaluate the results of
data analysis solution to inform the decision-making
process.

V0.0 Visuals Handout – Page 1

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.5 Data Visualisation Topic 9 - 9.6

Review Quiz Review Quiz

1. What is the primary purpose of model selection in 2. What is an example of a model evaluation
machine learning? technique used in machine learning?
a. Choosing the most complex model available
b. Selecting the model with the highest training a. Feature engineering
accuracy b. Cross-validation Correct Answer
c. Identifying the best-performing model for a specific c. Data preprocessing
task Correct Answer
d. Hyperparameter tuning
d. Picking a model randomly from the available
options

Data Visualisation Topic 9 - 9.7 Data Visualisation Topic 9 - 9.8

Review Quiz Review Quiz

3. In model evaluation, what does the term 4. Which of the following is a metric commonly used
"overfitting" refer to? for regression model evaluation?
a. The model performing well on the training data but poorly a. Precision
on new, unseen data Correct Answer b. Recall
b. The model being too simple and unable to capture the
underlying patterns in the data
c. Mean Squared Error (MSE) Correct Answer
c. The model's inability to generalize to different types of tasks d. Area Under the Receiver Operating Characteristic
d. The process of selecting the most complex model available (ROC) Curve

V0.0 Visuals Handout – Page 2

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.9 Data Visualisation Topic 9 - 9.10

Review Topic 8 Some Facts

• Model selection is a critical step because the choice of model can • Jupiter’s moons were first seen by Galileo in 1610 through a telescope.
significantly impact performance & accuracy of predictions or analyses. • Pasteur described disease processes around 1860 after seeing the
– Choose the method aligned with problem characteristics, goals, bacteria through a microscope.
and data availability. • Technologies such as radar, sonar, and medical scanners extend
• Model evaluation is the process of assessing the performance and human vision in powerful ways that facilitate understanding.
quality of a machine learning or statistical model. • Insights gained provide support for air traffic controllers, naval officers,
– Techniques and metrics to determine how well a model generalizes physicians and others in making timely and effective decisions.
to new, unseen data or accomplishes its intended task.
– Relies on specific performance metrics that are relevant to the
problem at hand.
– Metrics include accuracy, precision, recall, F1-score, mean squared
error (MSE), root mean squared error (RMSE), area under the
receiver operating characteristic curve (AUC-ROC), and many
others.

Data Visualisation Topic 9 - 9.11 Data Visualisation Topic 9 - 9.12

Some Quotes Visual Aids for Thinking

• “Visual perception of data is the visual means of resolving logical
problems” 5
• We build tools to magnify our cognition
– Jacques Bertin, one of the fathers of data visualization (1977)
• Example: multiplication (Card et al., 1999)
• “The greatest value of a picture is when it forces us to notice what  Multiply 35×95 (in your head)
we never expected to see”
– John Tukey, data scientist, inventor of the box plot visualization (1977)
 Now do it on paper
100
 People are 5 times faster with visual aid 95

• “Overview first, zoom and filter, then details-on-demand”

– Prof. Ben Shneiderman (1996)

• “I’ll Know It When I See It (IKIWISI)”

5 35
– Prof. Matthew Ward (2002)
40
SHNEIDERMAN, B. (1996). The Eyes Have It: A Task by Data Type Taxonomy for Information Visualizations. In Proc. IEEE Visual Languages 96. CARD, S.K., MACKINLAY, J.D., and SHNEIDERMAN, B. (1999). Readings in Information Visualization: Using Vision to Think. San Francisco,
WARD, M.O. (2002). A taxonomy of glyph placement strategies for multidimensional data visualization. Information Visualization (2002) 1, pp. 194– CA: Morgan Kaufmann.
210.

V0.0 Visuals Handout – Page 3

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.13 Data Visualisation Topic 9 - 9.14

What Can Visualization Do? Principles of Graphical Excellence

• Show the data
• Induce viewer to think about the data
• Present many numbers in a small space
Well-designed presentation of interesting data
• Reveal the data at many levels of detail, from  A matter of substance, of statistics, and of design
overview to fine structure
• Make a large dataset coherent  Consists of complex ideas communicated with clarity,
• Avoid distorting what the data have to say precision, and efficiency
• Encourage the eye to compare different pieces of
data
 Gives to the viewer the greatest number of ideas in the
• Serve a clear purpose (e.g., description,
shortest time with the least ink in the smallest space
exploration, tabulation)
 Requires telling the truth about the data
• Be closely integrated with the descriptions of a
dataset
TUFTE, E.R. (2001). The Visual Display of Quantitative Information. 2nd Edition. Cheshire, CT: Graphics Press.

Data Visualisation Topic 9 - 9.15 Data Visualisation Topic 9 - 9.16

What does Visualization Do? What is Data Visualization?

Often visualization focusses on three types of goals • Data is the fundamental element
• To explore driving the decisions across this
 Nothing is known about the data design process.
 Visualization used for data exploration • Without data there is no material to
 To find patterns from the data feed nor necessitate a visualization.
• To analyze • Conversely, without visualization
 There are hypotheses the value of data can be unfulfilled.
 Visualization used for Verification or Falsification • This is not to say we should always
• To present visualize data, absolutely not, but in
 “Everything” is known about the data most circumstances, to harness the
 Visualization used for Communication of Results maximum value of data, there are
missed opportunities if we do not.

V0.0 Visuals Handout – Page 4

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.17 Data Visualisation Topic 9 - 9.18

Why Data Visualization: Example Why Data Visualization: Example

• Company sales are presented What if you wanted to:
• Identify the first occasion when the
in Table: % share of online sales exceeded
 Find out what the percentage the % share of store sales?
share of online sales for a • When was the last occasion the %
Company X was during April share of store sales exceeded that of
2016? online sales?
• During which periods did the
different sales channels experience
Easy to Answer !
the most accelerated upward or
downward changes?

Difficult to Answer from current

representation!

Data Visualisation Topic 9 - 9.19 Data Visualisation Topic 9 - 9.20

Why Data Visualization: Example Why Data Visualization: Example

Visual presentation includes design choices such as the possible
• Figure shows same data. application of interactivity, features of annotation, all matters
• It is represented differently. around colour usage, and the composition of the work.
• Which is better depends
on the purpose  You could envisage interactivity being useful to offer tooltip
details of value labels as you hover over parts of each line.
 You could offer controls to modify the x-axis time range or
You do not plot data because filtering to hide or show different lines of interest.
you can, you do it because it  You could also add captions to provide explanations about
provides a window for seeing some of the most noticeable patterns in the data.
different features of data.  There are some features of annotation already on display with
this chart, such as the title, the colour legend, and the x- and
y-axis scales.

V0.0 Visuals Handout – Page 5

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.21 Data Visualisation Topic 9 - 9.22

Three Phases of Understanding:

Three Phases of Understanding Perceiving
Act of reading a chart: ‘what do
I see?’

 What chart is being used?

 What items of data do the
marks represent?
 What value associations do
the attributes represent?
 What range of values are
displayed?
 Are the data and its
representation trustworthy?
Kirk, A. (2016). Data visualisation: A handbook for data driven design. Sage.

Data Visualisation Topic 9 - 9.23 Data Visualisation Topic 9 - 9.24

Three Phases of Understanding: Three Phases of Understanding:

Perceiving Interpreting
Initial observations begin to form Translates observations into
about the main characteristics of the quantitative / qualitative meaning.
display: Assimilating what you observed
against what you know about the
 What features – shapes, patterns, subject.
differences or connections – are What does it mean, given the subject?
observable?
 Where are the largest, mid-sized
and smallest values? (known as  What features – shapes,
‘stepped magnitude’ judgements). patterns, differences or
 Where are the most and the connections – are interesting?
least? Where is the average or
normal? (‘global comparison’  What features are expected or
judgements). unexpected?
 What features are important
given the subject?

V0.0 Visuals Handout – Page 6

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.25 Data Visualisation Topic 9 - 9.26

Three Phases of Understanding:

Checkpoint Summary
Comprehending
Comprehending where viewers consider
what the interpretations mean to them. • Visualization is the process of representing data or information
graphically to make it more understandable, accessible, and meaningful.
 What can be inferred as being
important to you about the • It involves using visual elements such as charts, graphs, maps,
interpretations you have made? diagrams, and images to present complex data or concepts in a way
 What has been learnt? Has it that can be easily interpreted and analyzed by humans.
reinforced or challenged existing
knowledge? What feelings have been
stirred? Has the experience had an • Without visualization the value of data can be unfulfilled
impact emotionally?
 What does one do with this • This is not to say we should always visualize data, absolutely not, but in
understanding? Is it just knowledge most circumstances, to harness the maximum value of data, there are
acquired or something to inspire missed opportunities if we do not.
action, such as making a decision or
motivating a change in behavior?

Data Visualisation Topic 9 - 9.28

Visualization Types
• Linear/1D
• Planar/2D (incl. Geospatial)
• Temporal
• Volumetric/3D
• Multidimensional/nD
• Tree/Hierarchical
• Network
Introduction to Data Science and Big Data
Topic 9: Lecture 2 • Heatmaps
Types of Visualization • Infographics

© NCC Education Limited

V0.0 Visuals Handout – Page 7

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.29 Data Visualisation Topic 9 - 9.30

Linear/1D Linear/1D
Business Use Case: Stock price trends over time
• Lists of data items • Representing data along a single axis.
• Organized by a single feature (e.g., alphabetical order)
• Commonly visualized using bar charts, pie charts, etc.
• Excel tooltips explain quite well which visualization is
appropriate for which data.

Data Visualisation Topic 9 - 9.31 Data Visualisation Topic 9 - 9.32

Planar/2D Planar/2D

Types of Planar/2D Visualizations: Business Use Case: Sales

• Choropleth map Performance
• Planar/2D visualization is
• Cartogram map commonly used in business
• Dot distribution map to represent sales
performance data.
• Proportional symbol map • By plotting sales data on a
• Contour/isopleth/isarithmic map two-dimensional plane,
businesses can easily
• Dasymetric map compare the performance of
different products, regions,
• Self-organizing map or time periods.
Business Use Case:
Geographical distribution.

V0.0 Visuals Handout – Page 8

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.33 Data Visualisation Topic 9 - 9.34

Planar/2D Temporal
Types of Temporal Visualization:
Business Use Case: Geographical distribution. • Timeline
• Time Series
• Connected Scatter Plot
• Gantt Chart
• Stream Graph
• Arc Diagram
• Polar Area/Rose/Circumplex Chart
• Sankey Diagram
• Alluvial diagram
Proportional Symbol Map

Data Visualisation Topic 9 - 9.35 Data Visualisation Topic 9 - 9.36

Temporal Temporal
Business Use Case:
Monthly website Business Use
traffic fluctuations. Case: Project
• By visualizing sales Management
data over time,
businesses can
identify seasonal A Gantt chart
trends, understand representing the
the impact of scheduling and
marketing the dependency
campaigns, and (i.e., precedence
make informed network)
decisions for future relationships
sales strategies. between
activities.

V0.0 Visuals Handout – Page 9

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.37 Data Visualisation Topic 9 - 9.38

Volumetric/3D Visualization Multidimensional/nD Visualization

Business Use Case: Business Use Case:
Visualizing three- Analyzing customer
dimensional behavior with multiple
manufacturing variables.
processes.
Visualizing data with
Representing data in more than three
three-dimensional dimensions.
space.

Data Visualisation Topic 9 - 9.39 Data Visualisation Topic 9 - 9.40

Multidimensional/nD Visualization Waterfall Chart

Business Use Case:
Business Use Case: Product Income statement
comparison with multiple
variables.

Cumulative effect of
• Multivariate data sequentially introduced
• Three or more quantitative positive or negative
variables changes.
• Represented on axes from
the center Combination of bar
chart and candlestick
• Each line represents one chart
data point

Radar Chart or Spider Chart

V0.0 Visuals Handout – Page 10

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.41 Data Visualisation Topic 9 - 9.42

Tree or Hierarchical Visualization General Tree

Types of Hierarchical Visualization:

• General Tree Business Use
Case: Problem
• Dendrogram
solving
• Radial Tree
Organization -
• Hyperbolic Tree How items are
• Tree Map subdivided or
• Wedge Stack Graph (Radial Hierarchy) clustered.
• Icicle or Partition Chart

Data Visualisation Topic 9 - 9.43 Data Visualisation Topic 9 - 9.44

Icicle or Partition Chart Heatmaps

• Graphical representation
Emphasis on clusters, their size, and what level they of numerical data, where
are at. individual data points in
the dataset are
represented using
different colours.
• Heatmaps simplify
complex numerical data
into visualizations that
can be understood at a
glance.

V0.0 Visuals Handout – Page 11

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.45 Data Visualisation Topic 9 - 9.46

Heatmaps Example Infographic

• According to Oxford
Dictionary: “a visual
representation of
information or data”.

• A collection of imagery, data

visualizations like pie charts
and bar graphs, and minimal
text that gives an easy-to-
understand overview of a
topic.

Data Visualisation Topic 9 - 9.47 Data Visualisation Topic 9 - 9.48

Next Topic 10:

Topic Summary Business Intelligence and Tools
• Visualization is the process of representing data or information
graphically to make it more understandable, accessible, and meaningful. • What is Business Intelligence?
• Benefits of Business Intelligence
• Three Phases of Data Visualization: Perceiving, Interpreting,
Comprehending • Types of BI tools and software
• Example of BI case studies
• Visualization Types: Linear/1D, Planar/2D (incl. Geospatial), Temporal,
Volumetric/3D, Multidimensional, Tree/Hierarchical, Network,
Heatmaps, Infographics

• So many types of visualization, we have covered few, however, data

visualization can be done in different ways, and it depends on the
problem in hand which way to choose.

V0.0 Visuals Handout – Page 12

Topic X – Topic Title Module Title

Data Visualisation Topic 9 - 9.49

References

• Card, S.K., Mackinlay, J.D., & Shneiderman, B. (1999).

Readings in Information Visualization: Using Vision to Think.
San Francisco, CA: Morgan Kaufmann. Topic 9 – Data Visualization
• Kirk, A. (2016). Data visualization: A handbook for data
driven design. Sage.
• Tufte, E.R. (2001). The Visual Display of Quantitative Any Questions?
Information. 2nd Edition. Cheshire, CT: Graphics Press.

V0.0 Visuals Handout – Page 13