21AI602 - DATA VISUALIZATION

UNIT 1 – INTRODUCTION TO DATA VISUALIZATION

LP 3 – Data representation, Data Presentation, Seven stages of
data visualization, Widgets
1. Data Visualization Presentations:
 Data visualization is the art of converting contexts, circumstances, and developments
derived from verifiable data into visual objects in such a way that they are, ideally,
graspable and comprehensible right from the first glance.
 Infographics, pie charts, bar graphs, or radar charts are all popular formats.

Fig 1.1 Data Presentation

But how can larger volumes of data also be used to paint an informative picture? In five
steps, we’ll explain what you need to remember when creating charts, diagrams, and
graphs.

Step 1:
The data assessment

 Companies generally have access to a wide range of data. Be selective when choosing
the data you want to communicate.
 Assess the informative value and knowledge gain offered by the data sets. When it
comes to graphs, charts and diagrams, the same approach applies as for all other
PowerPoint-presentation content:
 Keep your presentation as short and concise as possible, and avoid overloading it with
superfluous information.
Data is relevant if, for example:

 It shows developments that enable forecasts to be made for the future.

 It describes pattern deviations that indicate changes in trends.
 It highlights contexts/correlations previously unheard of.
 It confirms previously unverifiable assumptions.

Step 2:
The message

 Data is often multidimensional, which makes it complex and difficult to understand.

When it comes to data visualization, it is important to concentrate on the main findings
and make clear, simple statements.
 But this also means you need to leave out anything that’s unnecessary. Excessive detail
does not make a chart better.
 A core statement is not made more concise by having extra aspects added on to it. When
handling data, people often fall into the trap of following an almost scientific approach.
 But most business presentations are not of a scientific nature; they’re about key
findings. And the more scaled back these are, the clearer they become.

Step 3:
The target audience

 Carefully consider the target audience of your presentation, and think about how
familiar they are with handling data.
 While data is seen as proof of certain statements, it often also raises new questions.
 Always state the data source, and be prepared for questions about the collection method,
time frame, and contexts/correlations.

Step 4:
Navigation

 When visualizing your data, only apply principles that will help your audience navigate
and grasp the message.
 Making a table colorful because it looks nicer is not constructive; it might even end up
being confusing.
 The human brain takes in lots of information subconsciously and sorts it. And there are
certain perception principles you can utilize.
 Give your audience whatever they need to understand and easily grasp what you are
showing, such as:
 A clear verbal introduction
 A clear heading conveying the slide’s core statement
  Colors of emotional significance (red=danger, yellow=neutral, green=desirable,
corporate colors and competitor colors)
 Logical reading order (left to right and clockwise, e.g. based on importance and
percentage in a pie chart)
 Labelling and accompanying texts reduced to the absolutely necessary, avoid repetition
 Learned symbols, e.g. symbols for female and male or flags
 Important information highlighted—using size, color, or distinguishing elements

Step 5:
Implementation

 Only once you are familiar with your data and have identified the main statement can
you determine the type of data suitable.
 It’s not just the visual attractiveness of a display format that is crucial; its function is
too. Situations can be shown differently to processes.
 Contexts/correlations require a different format to percentages.
 A list of the most common display formats and their areas of use is available here.

Fig 1.2 Different types presentation

2. Data Visualization Representation

 Effective data visualization involves thoughtful consideration of various elements to
represent information accurately and facilitate understanding.

Choosing the Right Chart Type:

 Selecting an appropriate chart type is crucial for conveying the intended message.
  Bar charts, line graphs, pie charts, scatter plots, and heatmaps are just a few examples
of different chart types, each suited for specific types of data.

Color Coding:

 Color can be used to highlight important information, differentiate data categories, and
draw attention to specific elements.
 Careful color choices enhance clarity and readability, while misuse can lead to
confusion.

Symbolism:

 Using symbols or icons can help convey information more intuitively.

 For example, icons representing different types of data points or shapes indicating
specific categories can make the visualization more accessible.

Size and Proportion:

 Utilizing size and proportion in elements like bars or bubbles can represent quantitative
values effectively.
 Larger sizes often indicate higher values, allowing for a quick visual comparison.

Spatial Arrangements:

 The arrangement of elements on a visual display can influence the viewer's

interpretation.
 Grouping related items, arranging data chronologically, or organizing data
hierarchically can enhance comprehension.

Animations:

 Animations can be used to reveal trends over time, transitions between states, or
changes in data.
 However, animations should be employed judiciously to avoid distracting the audience
from the main message.

Patterns and Trends:

 Visualizations should highlight patterns and trends in the data.

 This might involve connecting data points, using trend lines, or employing other visual
elements that emphasize the narrative or insights.

Data Density:

 Data density refers to the amount of information presented in a given space.

 Striking the right balance is essential; too much information can be overwhelming,
while too little may not provide enough context.
  Techniques such as small multiples can help manage data density effectively.

3. Seven stages of Visualizing Data Process

 The Ben Fry visualizing data process is very helpful. It involves 7 steps. For more
details, refer to his book: Visualizing Data.
1. Acquire: Obtain the data, whether from a file on a disk or a source over a network.

 Data can be collected from many resources such as books, files and digital documents.
This is the beginning and fundamental step of data visualization.
2. Parse: Provide some structure for the data's meaning, and order it into categories.

 You may have collected immerse data, but it is necessary to restructure the collected
data.
 This structure will make it easier to know convey to others what data you have by
format, tags, names, and indices.
3. Filter: Remove all but the data of interest.

 Not all data is useful. Filter out the data that cannot serve your goal.
 If you are focusing on the data of a specific period, remove the data of other periods.

4. Mine: Apply methods from statistics or data mining as a way to discern patterns or place the
data in mathematical context.

 Data visualization is to help viewers seek for insights that may not be gained from raw
data or statistics.
 This step helps get basic understanding of the data that is significant for the whole
process.
5. Represent: Choose a basic visual model, such as a bar graph, list, or tree.

 Various visual models are available.

 You need to select the suitable type that best suits your need.
 Edraw data visualization software helps you produce over 200 kinds of visuals instantly.
6. Refine: Improve the basic representation to make it clearer and more visually engaging.

 Polish your work according to some basic color and graphic design theory.
 You can also rely on Edraw by just choosing one theme which includes harmonious
color, font, and line style.
7. Interact: Add methods for manipulating the data or controlling what features are visible.
4. Widgets:
 A widget displays the results of a dataset, allowing you to visualize data in different
ways than traditional reports displayed as grids or graphs do.
 Widgets are sophisticated visualization techniques that can combine with rich
interactivity to enable users to understand their data more effectively.
 Dossier-style documents can display a variety of widget types, such as Gauge, Heat
Map, and Stacked Area widgets.
 Although each type of widget looks different and is used in a unique way, the main
purpose of all widgets remains the same: to provide a visual and interactive look into
your data.

Fig 4.1 Widgets

To analyze data using a specific type of widget

i. Bubble Grid widget: Displays bubbles of different colors and sizes representing
the values of two metrics. It can help identify important trends or anomalies in data,
relative to the total contribution of accompanying data.

ii. Cylinder widget: A simple status indicator that displays a vertical cylinder with
fluid in it. The level of the fluid within the cylinder is a visual representation of a
single metric value. This is a Flash-based widget, which displays in Flash Mode.
iii. Data Cloud widget: A list of attribute elements displayed in various sizes to depict
the differences in metric values between the elements. The varying sizes allow you
to quickly identify the most significant, positive, or negative contributions.

iv. Date Selection widget: A calendar selector that allows you to select which dates
you want to see data about in a document or dossier-style document. You are able
to see all of the dates of each month in the widget, which allows you to be able to
select dates more easily.

v. Fish Eye Selector widget: An interactive selector that magnifies an item when you
hover the cursor over it. It allows you to choose from a list of attribute elements,
metrics, or images without having to see all of the elements, metrics, or images.
Any item that you hover over or select remains magnified, while the remaining
items are minimized and hidden from view. This can be especially helpful when
you have to browse through a lengthy list.

vi. Funnel widget: A variation of a stacked bar graph that displays data that adds up to
100%. It allows you to visualize the percent contribution of a metric to the whole.

vii. Gauge widget: A simple status indicator that displays a needle that moves within a
range of numbers displayed on its outside edges. An example of a gauge is a car's
speedometer.

viii. Graph Matrix widget: A group of area graphs that display actual values and line
graphs that display forecasted values. It allows you to quickly analyze various
trends across several metric dimensions.

ix. Heat Map widget: A combination of colored rectangles, each representing an

attribute element, that allow you to quickly grasp the state and impact of a large
number of variables at the same time.

x. Image Layout widget: An image overlaid with colored areas or bubble markers.
For example, the layout of a store can be displayed, with each aisle shown as a
separate region. Each aisle is automatically colored based on the number of visits
that it receives.

xi. Interactive Bubble Graph widget: A conventional bubble plot that allows you to
visualize the trends of three different metrics for a set of attribute elements.
xii. Interactive Stacked Graph widget: A combination of a checkbox list and area
graph. The graph allows you to see the contribution of various metric series to the
change in value of a larger set of data. This is a Flash-based widget, which displays
in Flash Mode.

xiii. Map widget: Locations displayed as image markers or bubble markers on a map,
along with additional data for those locations, such as attribute and metric data.

xiv. Media widget: Video, audio, images, or website content. One of the primary
purposes of the Media widget is to present supplemental information about the data
on a dossier-style document. It can also be used for instructional content or HTML
content from a website.

xv. Microcharts widget: One or more compact representations of data that allow you
to quickly visualize trends. Use a Microcharts widget to quickly visualize the trend
of a metric at a glance without having to know many additional details. The bar,
sparkline, and bullet microcharts used in the Microcharts widget convey
information that you can understand just by looking at the graph once.

xvi. Network widget: A visual representation of data about individual items and the
relationships between them. Business attributes are represented by circular nodes,
while the lines between the nodes (called edges) represent relationships between the
attributes.

xvii. RSS Reader widget: RSS (Really Simple Syndication) is a data format used to
display updated Web content when you click a URL. An RSS document is called a
feed, and it contains either a summary of the content from an associated website or
the full text. The RSS Reader widget can help provide context to your business data.
Use RSS Reader widgets on a dossier-style document to view and update RSS feeds
while you analyze grids, graphs, and other objects in the same dossier-style
document.

xviii. Thermometer widget: A simple status indicator that displays a thermometer set to
a certain temperature level. The temperature level within the thermometer is a visual
representation of a single metric value. This is a Flash-based widget, which displays
in Flash Mode.

xix. Time Series Slider widget: An area graph that allows you to choose which section
of the graph to view at a time. This is a Flash-based widget, which displays in Flash
Mode.

xx. Waterfall widget: A group of clustered bars displayed from left to right. It
highlights the increments and decrements of the values of metrics over time. The
 widget can help identify what is contributing to fluctuations in the metric values
and can be used for "what-if" analyses.

xxi. Weighted List Viewer widget: A combination of the data visualization techniques
of thresholds and graphical weighting in a single visualization. This enables you to
assess the performance of a group of items.