Introduction to Data Analytics
Data Visualization for Decision-Making
Learning Objectives
By the end of this lesson, you will be able to:
Explain data visualization
Describe the importance of data visualization
List various tools of data visualization
Data Visualization
Data Visualization
● Data visualization is the graphical representation of
data using charts, graphs, and maps.
● Our eyes are drawn to colors and patterns.
● Data visualization is a form of visual art that grabs
our interest and keeps our eyes on the message.
Data Visualization
Visualized data is more effective and consumable than a massive
spreadsheet of data.
Data Visualization
● The table shows the total sale of products
corresponding with each year.
● The adjacent graph is the visualization
of the sale data points using a
visualization tool.
Understanding Data Visualization
Understanding Data Visualization
Data analytics allows decision makers and executives to weigh the alternatives of different outcomes of
their decisions.
Helps decision makers strategize Provides answers to key
the best business outcomes Data Visualization business questions
Provides simplicity, clarity, intuitiveness, insightfulness, pattern,
and trending capability to help executives take decisions
Benefits of Data Visualization
● Sales reports are formal documents or
PowerPoint slides with many tables and charts.
● They are elaborate and the real point is lost in
the data.
● Data visualization helps by making information
crisp, clear, and memorable.
Benefits of Data Visualization: Example
● From a bar chart, a sales director can
identify that the sales of their flagship
product in the southwest region is going
down by eight percent.
● The director can spot the occurrence of
variances and start formulating a plan to
improve the sale.
● Data visualization allows executives to
spot problems and act on them.
Data Visualization: Uber
In October 2018, Uber released a visualization product that provides insights on mobility for JUMP
bikes.
This data is helpful in urban areas to evaluate the success of their shared bike program.
It can also help in planning infrastructure investments in cities to promote safety and smooth mobility.
Commonly Used Visualizations
Commonly Used Visualizations
Frequency
Heat map Distribution Swarm Plot
Plot
Heat map
Heat map
● A heat map is a type of graph that uses a warm-
to-cool color spectrum to visualize the data.
● It measures the relationship between multiple
variables and shows the strength of
relationships with colors.
● It helps in creating a visually impactful view of
correlation.
Heat map: Use Case
1 10 11 12 12 14 15 100
● All the rows are one category and all the 16 17 18 19 20 21
columns are another category. 2
● Individual rows and columns are divided 3 50
into subcategories.
4
● Cells either contain color-coded
categorical data or numerical data. 5 0
Row A B C D E F
name
Column name
Heat map: Use Case
1 10 11 12 12 14 15 100
● Data in a cell is based on the relationship
16 17 18 19 20 21
between two variables in the connecting 2
row and column.
3 50
● Multiple value ranges can be
represented by a selection of solid 4
colors, while a single range can be shown
by a gradient scale. 0
5
Row A B C D E F
name
Column name
Heat map: Case Study
This heat map shows the sales data across
months and year.
● It is observed that during 1949–1950,
sales were in range of 0–200.
● Sales crossed the 200 mark in 1951 and
increased every year.
● In 1956, sales were between 300 to 500
and the maximum sale was in July.
● Over the years, July had the maximum
sales and it peaked in 1960.
Frequency Distribution Plot
Frequency Distribution Plot
Frequency distribution plot measures the frequency of occurrence for a given value or range.
A normalized frequency distribution normalizes total frequency to one.
The frequency of an event is the number of times the event occurs in an observation.
Observations within a given interval are in graphical or tabular format.
Frequency Distribution Plot
Frequency distributions can be displayed in these formats:
Table Histogram Line graph
Dot plot Pie chart
Analysts use the frequency distribution plot to check or illustrate the data collected in a
sample.
Frequency Distribution Plot: Use Case
This conditional frequency distribution graph shows the usage of two specific words in public
speeches over a period of time.
The graph indicates that the use of America has increased, while the use of citizen has reduced
gradually in speeches.
Swarm Plot
Swarm Plot
● A swarm plot gives a good representation of
the distributions but works well only for small
data sets.
● It is useful to examine individuals, places, or
things in your data.
● It allows you to plot all of your points in a
single space.
● It is a one-dimensional scatter plot as it plots
the data on a single axis and then offsets in
the other direction to show volume.
Swarm Plot
A swarm plot enables you to separate all overlapping points, making each point visible.
Beeswarm plot
It is also called a beeswarm plot as the graphical representation is similar to a group of bees.
Swarm Plot: Use Case
In this swarm plot:
● The X-axis denotes time and the Y-axis denotes
the tip amount
● The blue graph represents lunch and the
orange graph represents dinner
● Tips are higher during dinner and most tips are
$2 and $3
● The maximum tip is $10, which was given for
dinner.
Importance of Data Visualization
Importance of Data Visualization in Analytics
● Data visualization tools provide access to
trends, outliers, and patterns in data.
● They help organize and present important
findings from the data.
● Data in user-friendly charts help businesses
gain insights to make right decisions.
Importance of Data Visualization in Analytics
● Data analytics tool allows a user to present
massive data intuitively.
● Decision makers see patterns, trends, and
correlations in the data being analyzed.
● It help decision makers in cutting costs or
improving operational processes.
Exploratory Data Analytics
Exploratory Data Analytics
● Exploratory data analytics is an approach to
analyze data sets to summarize their main
characteristics.
● Data visualization in exploratory data analytics
is the first step towards modeling.
● EDA primarily helps analyze data beyond the
formal modeling.
Exploratory Data Analytics
The steps involved in EDA are:
● Get detailed insights into the dataset
● Understand critical impact variables that
influence the dataset
● Detect if any outliers are present in the
dataset
● Test the underlying assumptions of the
dataset
Data Visualization Tools
Data Visualization Tools
● Tableau is the most widely used data
visualization tool due to its simplicity and
ability to produce interactive visualizations.
● It has a large customer base of more than
50,000 accounts across many industries.
● FusionCharts is a JavaScript-based visualization
package that can produce and integrate 90
different chart types.
● It has a range of live example templates so that
you can simply plug in your data sources as
required.
Data Visualization Tools
● Highcharts is often chosen when a fast and
flexible solution has to be rolled out.
● Its cross-browser support feature helps users
view and run interactive visualizations.
● Datawrapper has a simple interface that makes
it easy to upload CSV data, create charts, and
maps.
● It is becoming popular among media
organizations to create charts and present
statistics.
Data Visualization Tools
● Plotly enables complex and sophisticated
visualizations.
● It is integrated with analytics-oriented
programming languages such as Python, R, and
Matlab.
● Sisense provides a full stack analytics platform
and simple-to-use drag and drop interface.
● It helps in creating charts and complex graphics
with minimum hassle and provides a repository
for gathering multiple sources of data.
Other Visualization Tools
Other Visualization Tools
● Power BI is a powerful suite of business analytics
tools and has intuitive UI for users familiar with
Microsoft products.
● It can create customized, user-defined
visualizations as well as sophisticated 3D maps.
Power BI
● The Looker BI tool provides extensive
visualization abilities, along with real-time
analysis.
● Users can either use templates from the Looker
library or create a custom visualization.
Other Visualization Tools
● Domo is self-service business intelligence
that focuses on social collaboration.
● It provides real-time data and uses creative
data displays such as multi-part widgets
and sparklines.
● Board is a full-featured business
intelligence system.
● It serves midsize and enterprise-level
companies in different industry segments.
Other Visualization Tools
● Qlik Sense has a clean and clutter-free user
interface and a highly customizable setup.
● Qlik Sense is Tableau’s biggest competitor.
● It has over 40,000 customer accounts in more
than 100 countries.
Languages and Libraries for Data Visualization
Languages for Data Visualization
Few languages and libraries leveraged by
data visualization:
● Scala
● R
● Python
● Javascript
● Java
Languages for Data Visualization
Python has two exclusive libraries for
data visualizations that are
Matplotlib and Seaborn.
Scala is a compiled language and
the code written in Scala gets
executed much faster.
Languages for Data Visualization
Base graphics, lattice graphics, grid
graphics, and ggplot2 are the four
graphic systems supported by R.
Inbuilt libraries available in Java such
as Java 2D, Java 3D, and Java
advanced imaging makes data
visualization simple with Java.
Java
Java 2D Java 3D
Advanced
Data Visualization Libraries
Shiny is an R Matplotlib is the Seaborn is a Python Bokeh is native of
package and it is first data data visualization Python and helps to
easy to build visualization library library and provides create interactive,
interactive web and has 2D and 3D an interface for web-ready plots by
apps straight from graphics support. drawing statistical supporting
R. graphics. streaming and real-
time data.
Dashboard-Based Visualization
Dashboard-Based Visualization
Provides a visual real-time representation Displays data in the form of tables, line
of a company’s data. charts, bar charts, and gauges.
Business
Dashboard
Monitors business health by visually
Helps businesses generate business
tracking, analyzing, and displaying key
insights.
data points.
Dashboard-Based Visualization
This dashboard has parameters such as daily target, sales pattern, and
other business insights from different charts.
Characteristics of effective dashboards:
● Highly interactive
● Customizable interface
● Pulls real-time data from multiple
sources
Steps for Dashboard-Based Visualization
● Analyze your target audience
● Identify Key business parameters
● Identify the end goal of the dashboard
● Get hands-on in developing the dashboard
● Continuous process of improvement
Steps for Dashboard-Based Visualization
Analyze your target
audience Know who will use the data to make decisions.
Identify key business
Identify KRAs and KPIs for each Key Process Area (KPA) and
parameters
Service Level Agreement (SLA) parameters.
Identify end goal of the
Define a target dashboard outline.
dashboard
Get hands-on in Help data scientists develop the dashboard by selecting
developing dashboard convenient languages and libraries.
Continuous process of Improve the dashboard based on real-time inputs and customer
improvement feedback.
BI and Visualization Trends
BI and Visualization Trends
● The development of BI to analyze and
extract value from various sources
introduced many errors and low-
quality reports.
● Companies choose to implement the
Data Quality Management (DQM)
policy as it is a key factor to efficient
data analytics.
BARC Research Study
A survey was conducted by the Business
Application Research Center (BARC) on the BI
trends.
Users
Users, consultants, and vendors were among
the 2865 participants for the survey.
2865
Master data/Data quality management was Consultants
participants
stated as the most important trend in 2020.
Vendors
BARC Research Study
● BI practitioners identified Master
data/DQ management, Data discovery,
and Data-driven culture as the three
most important trends in their work.
● Cloud BI/data management, Data catalogs,
and Process mining were voted as the
least important trends.
BARC Research Study
● Master Data and Data Quality
Management build a strong foundation
for handling data.
● Data Discovery describes how
businesses collect data from various
sources and then apply it to generate
real business value.
BARC Research Study
● The significance of Data discovery
shows a strong trend in business
users empowerment.
● Data-driven culture depends on
greater inclusion of various business
departments.
● Data governance is an important
trend due to the GDPR and increase
in data security awareness.
BARC Research Study
Trends that have increased in importance
compared to the last year are:
● Data-driven culture
● Real-time analytics
● Integrated platforms for BI and PM
● Embedded BI
● Analytics Team/Data labs
BARC Research Study
● In Data-driven culture, all decisions and
processes are based on data and simple
key figures like revenue.
● Real-time analytics is about capturing
new data immediately after their
occurrence and processing them for
display or analysis.
● Embedding intelligence is growing its
popularity in operational applications.
BARC Research Study
● Mobile BI has grown by only 20 percent
in last eight years as the adoption was
very slow.
● Data catalogs and Process mining trends
are a little new in the market and has
only recently been attracting more
interest.
BI Software Challenges
BI Software Challenges
● A survey by BI_Survey.com indicated that
data quality is the top most problem for
BI software users since 2002.
● BI software output is largely and highly
impacted if the input data quality is not
good.
BI Software Challenges
Poor quality input data leads to bad output results provided by any BI tool.
BI companies compile data in a usable system without affecting the validity
of the original source.
Key Takeaways
Data visualization refers to the graphical representation of data
using charts, graphs, and maps.
Heat map, frequency distribution plot, and swarm plot are
commonly used visualizations.
Tableau, PowerBI, Datawrapper, and Sisense are some of the
data visualization tools.
