Unit-3
Classification of visualization systems
Classification of visualization systems:
• Common Types of Data Visualization
• Advanced Types of Data Visualization
Common Types of Data Visualization
1. Bar Graphs
In a bar graph, each bar's height
corresponds to the value it displays.
Each bar's height corresponds to the
numerical value it displays.
When to Use:Bar graphs are excellent for comparing numbers that fall into distinct categories
or groupings, such as
sales by location or students by grade.
Examples:A bar chart might display the monthly sales of a car lot or the income earned
by several product lines.
2. Line Graphs
To
show
values across time or other continuous variables, a line graph draws lines between
data points.
When to Use: Data trends or patterns over time or other continuous variables may be effectively
shown using line graphs.
Examples: A line graph might display weekly temperature swings or annual sales growth for a
business.
3. Pie Charts
A pie chart, often known as a circular bar graph with equal segments or "slices," is a graphical
representation of quantitative data.
When to Use: Percentages and other fractions, such as the
market share of various brands,
may be shown
clearly and concisely using pie charts.
Examples: A pie chart might be used to illustrate the breakdown of sales by
department or the breakdown of
the population by age bracket.
4. Scatter Plots
Each dot in a scatter plot represents the value of two variables, and the plot itself is a
representation of the relationship between those variables.
• When to Use: Using a scatter plot, one may see the association between temperature and ice cream
sales, for example.
• Examples: The link between unemployment rates and inflation, for example, or the relationship
between the height and weight of pupils at a certain school, may be visualized using a scatter plot.
5. Heat Maps
A heat map is a kind of bar graph in which different colors are used to denote different values.
The colors are often organized on a scale from dark to light.
When to Use: Heat maps are effective for visualizing trends and patterns in massive datasets,
such as those
describing website traffic or consumer de mographics.
Examples: Crime statistics throughout a city, or website traffic patterns, might
both be visualized using a heat map.
Advanced Types of Data Visualization Tree
maps:
• Treemaps are visualizations for hierarchical data. They are made of a series of nested
rectangles of sizes proportional to the corresponding data value. A large rectangle
represents a branch of a data tree, and it is subdivided into smaller rectangles that
represent the size of each node within that branch.
• When to Use: Tree maps are an effective way to display the structure and relative sizes of
several classes or groupings, such as product lines or market shares.
• Treemaps are particularly useful for visualizing the distribution of advertising spend in •
Network administrators often use
treemaps to anal yze disk space use. The rectangles
2.
Network Diagrams
• Examples: A tree map might display the distribution of various sorts of crime across a city,
for example, or the distribution of a company's money among its many divisions. a
diverse mix of digital and traditional ad channels.
are generally different colors to make differentiating each data set simple.
• Network Visualization (also called Network Graph) is often used to visualize complex
relationships between a huge amount of [Link] type of visualization illuminates
relationships between entities. Entities are displayed as round nodes and lines show the
relationships between them.
• The two notable types of Network Diagrams are "undirected" and "directed". Undirected
Network Diagrams only display the connections between entities, while directed Network
Diagrams show if the connections are one-way or two-way through small arrows.
• When to
Use:Social networks, transportation networks, and supply chains are all
examples of systems that benefit greatly from network visualizations.
• Examples:Social media connections, intercity travel patterns, and other data may all
be represented visually using
network diagrams
3. Chord Diagrams
• This type of diagram visualises the inter-relationships between entities. The connections
between entities are used to display that they share something in common. This makes
Chord Diagrams ideal for comparing the similarities within a dataset or between different
groups of data. Each arc's breadth reflects the depth of the bond.
• Nodes are arranged along a circle, with the relationships between points connected to each
other either through the use of arcs or Bézier curves. Values are assigned to each
connection, which is represented proportionally by the size of each arc.
• Colour can be used to group the data into different categories, which aids in making
comparisons and distinguishing groups.
• When to Use: Chord diagrams are helpful for depicting the interdependencies
and interconnections between various entities, such as the transfer of resources or
information between nations or businesses.
• Examples: Chord diagrams may be used to display everything from international
commercial ties to internal organisational networks.
4. Sankey Diagrams
• A Sankey diagram typically displays a flow of data with the help of links and nodes. The
nodes represent the entities or categories in the chart and can be sources (usually on the
left), destinations (usually on the right), or intermediate steps in a process. Several entities
(nodes) are represented by rectangles or text. The links are the connections between
nodes. The width of each link represents the volume of flow between each node.
• When to Use: The flow of energy in a power system, flow of money in a financial system
or the distribution of water resources, are two examples of situations where a Sankey
diagram may be helpful.
• Examples: The flow of
money in a financial system or the breakdown of GHG
emissions by industry might both be shown using
a Sankey diagram.
• Here is an example displaying the number of people migrating from one country (left)
to another (right). Data used comes from this scientific publication.
5. Radar Charts
• Radar Charts are a way of comparing multiple quantitative variables. This makes them
useful for seeing which variables have similar values or if there are any outliers amongst
each variable. Radar Charts are also useful for seeing which variables are scoring high or
low within a dataset, making them suited for displaying performance.
• Each variable is provided with an axis that starts from the centre. All axes are arranged
radially, with equal distances between each other, while maintaining the same scale
between all axes.
• Grid lines that connect from axis to axis are often used as a guide. Each variable value is
plotted along an individual axis and all the variables in a dataset and connected to
form a polygon.
• When to Use: Radar charts are helpful for comparing the values of many
distinct factors, such the effectiveness of various items or the
proficiency of various personnel.
• Examples: A radar chart may be used to compare the success of several movies along
many dimensions or to track the progress of a sports team across multiple metrics.
Misleading Data Visualization
Data visualization is a critical aspect of data analysis, as it helps organizations make sense of
large amounts of data and gain insights that are not immediately obvious.
However, data visualization can also be misleading if not done [Link] data
visualizations can lead to incorrect conclusions, misinterpretations, and ultimately, poor decision
making.
Understanding the factors that contribute to misleading data visualizations is critical for
organizations that want to gain meaningful insights from their data and make informed
decisions.
When properly implemented, data visualization has the ability to convey complex
information in a way that is easy to understand. While it is true that “a picture is worth a
thousand words,” this adage can quickly turn sour if a false image, chart, or graph causes
people to draw inaccurate
conclusions, make
the wrong decisions, and lose trust in a brand or organization.
Take, for instance, company X which provides a graph to indicate how its profits have
changed in the period of 5 months.
The company’s profits are 20.1M, 20.2M, 20.2M, 20.3M, and 20.4M for the first 5 months
respectively. The graph above has the y-axis starting at 20 instead of 0, making this look like
there was such a huge profit growth. The increase shows fake rises and can mislead users,
stakeholders, and other interested parties, and it may cause them to draw inaccurate conclusions
(although, technically speaking, the data shown is correct). This is a good example of a
misleading data visualization that might alter people’s perceptions. Now, let’s take a look at the
second graph below.
In this second graph, the y -axis starts at 0 and shows that the profit has
hardly changed. This graph can be accurately
represented like this as it will not exaggerate the data
that you want to show.
Misleading data visualization examples
– things to
avoid
Looking at bad data visualization exa mples is the best way to learn from them and avoid
including their flaws in your own product’s data visualization.
The following examples illustrate the kinds of things you should aim to avoid.
Truncating y-axis in graphs
A very common misleading data visualization example is changing the value of the y
-
axis’s starting point from zero to any other number. This blows up the differences when
comparing data. Let’s have a look at the graph below to see how a truncated graph might distort
data.
From the graph above, all data is in the 9300-9800 range on the y-axis and this tends to show an
exaggerated growth. This data can be much more accurately presented in this way:
Now we can see that the differences between the data of the different groups
are small . This is a more accurate picture of the differences and a fair way to present the
information.
Extending labels on the y-axis
A truncated Y -axis is a common mistake in data visualization where the
scale of the Y -axis is
artificially shortened to make changes in the data appears more significant.
It
This can lead to misleading visualizations and incorrect conclusions. can hide
the trends we’re trying to show, which can alter the magnitude of a change. In the graph below,
we can see the average
yearly global temperature from 1880 to 2015.
This graph is based on NASA data and it displays annual average temperatures dating back to
1880. The issue with this graph is that putting data on the extended y-axis makes it impossible to
spot how temperature really changes over the years. This is a common misleading data
visualization example. Now, let’s take a look at the second graph.
The graph above demonstrates how, during the same period of time, average temperatures
all around the globe have been gradually
increasing. This fits nicely with the predictions of the planet’s warming made by scientific
models. There is, of cour se, no use in displaying temperature labels that range from -110 to
110 degrees because the average temperature in the
recorded period never went below 56 and above 59 degrees Fahrenheit.
Exaggerated or improper scaling
Some people like to use images (pic tograms)
in their charts to make the visualizations
more interesting to the eye. This can look appealing but, if not done properly can be a bit
misleading. Consistent scaling is required for pictorial bar graphs. When the scale is off,
misle
it’s simple to
ad the audience. Let’s
take a look at the graph below:
While glancing at the picture on the graphs above, it is difficult to read the data correctly. While
the value for column B is indeed 3 times higher than for A, the size of the pictogram makes it
look
like the difference is much bigger. This graph can be better represented by using a bar chart
which is common and easier to read, or by stacking the images on each other as seen below
Improper extraction
– Cherry Picking
Cherry-picking data is the act of selecting only the data that supports a desired conclusion while
ignoring or downplaying data that contradicts it.
This is a common mistake in data visualization and can lead to
misleading visualizations and incorrect conclusions.
It is important to consider the context and limitations of the data when
creating a visualization.
By selecting just certain pieces of information to back up an argument rather
than the whole body of evidence, cherry -picking creates the false
impression that the story be ing told is more credible than it really is. Most of the time, details
that should have been in the graph are knowingly omitted.
The graph below is a good misleading data visualization example. This graph shows the average
global temperature from 1997 to 2012. The goal of this
graph is to prove that global warming is
not happening.
Interestingly, the graph suggests that the temperature has been quite stable. In constructing this
graph, the authors paid little attention to the 100 years before the time period in question; instead,
they focused only on events that gave credence to their claims. Just a small fraction of the
information was utilized, so it’s easy to understand how it may mislead readers. Below is how
this graph should have been presented:
The following chart shows the average temperatures from 1900 to 2020,
showing a clear increase in global temperatures.
Going against the norm
– unusual coloring
Good outcomes (earnings, victories, gains, etc.) are typically associated with the color
green, while bad outcomes (losses, etc.) are always associated with the color red. If you
deviate from these standard procedures, you increase the likelihood of creating
misunderstanding and maybe even
intentional distortions of the facts.
Here’s a chart that shows the state-by-state gains and losses. From this chart, using the color red
at first glance makes it look like the states are running at a loss which is not the case. A closer
look at the graph you’ll see that the states with the red color do not have negative numbers, while
the states with the green
color have negative numbers.
This is a not-too-common misleading data visualization example but it happens every now and
then. Let’s observe the second graph below.
Using
the color green makes it easy t o observe that
the number of profitable states far outweighs
the number of those losing money. Looking at the graph, the states with green don’t have
negative numbers, while the states with red have negative numbers. We can
see that changing standards such as color or order can have a significant
impact on how that information is perceived. Data visualizations are meant to swiftly
relay information. So stick to the norms to keep things simple.
3D graphs pie chart
A pie chart should always add up to
100%. H owever, making a pie chart 3D
or adding a slant
will make interpretation difficult due to the distorted effect of perspective.
On the right-most chart, the A and B values look much bigger than they actually are. This is
because values visually closer to the viewer will always appear oversized. This shows how
common it is for 3D pie charts to misrepresent angles. 3D pie charts have a tendency to throw off
proportions and make things appear larger or smaller depending on their angle; therefore, it’s
better to avoid using it.
Impact of misleading data visualization
Misleading data visualization might lead to erroneous conclusions and poor business choices that
may not be in the best interest of the company and that may have effects on your business. The
following are some of the impacts of misleading data visualization:
• Lead to poor choices based on false information
• Loss of growth opportunities
• Lose the trust of co-workers
• Miss out on trends
How to Avoid Misleading Visualizations
1. Maintain Accurate Scales: Always use proportional, consistent scales, especially for the
axes.
2. Choose the Right Chart Type: Make sure the chart type is appropriate for the data. For
example, use line graphs for continuous data and bar charts for categorical data. 3. Use
Transparent Context: Provide full context about the data, such as its source, the time
period, and any assumptions made.
4. Simplify Complex Visuals: Avoid overcomplicating visualizations with excessive 3D
effects, non-standard charts, or confusing color schemes.
5. Avoid Data Compression: Be cautious when truncating axes or compressing data ranges
to avoid distorting the story the data is telling.
Visualization of one, two and multi-dimensional data
Visualization is an essential component of data science, as it can help to communicate findings,
make predictions, and provide actionable insights.
1D Data Visualization
1D data visualization involves representing a single set of data values (usually along a single
axis) in a way that makes patterns, trends, or distributions easy to understand. Since the data is
one dimensional, it's typically displayed along a single line or axis, with points, bars, or lines
illustrating the values.
Here are some common methods for visualizing 1D data:
Histogram
Distribution of data values across different ranges (bins).Used when you want to see the
frequency distribution of continuous or discrete numerical data. Example: Showing the
distribution of test scores in a class
Bar Plot (Vertical or Horizontal)
Discrete values or categories with corresponding [Link] for comparing the frequency
or count of discrete categories or items.
Example :
Number
of occurrences of different categories (e.g., colors of cars in a parking lot).
Line Plot
The trend of data over time or across a [Link] for showing temporal or sequential data
(e.g., time series).
Example: Stock market prices over time, or temperature changes over days.
Scatter Plot (for 1D data with
repeated measurements)
Each data point plotted along the [Link] to visualize the
relationship or spread of data, often
with
repeated measurements or individual points along the 1D axis.
Example :
A set of individual test scores plotted along an axis.
2D Data Visualization
2D data visualization involves plotting data in two dimensions, typically using two numerical
variables along the x-axis and y-axis. It allows you to explore relationships between two
variables, trends, distributions, and patterns in a dataset. There are various ways to represent 2D
data depending on the type of data and what you're trying to convey.
Scatter Plot
Used to observe correlations, trends, or patterns between two variables. It is particularly useful
for identifying clusters, outliers, and linear/nonlinear relationships.
Example: Comparing height vs. weight, or hours studied vs. exam scores
A grid of data values with color indicating the magnitude of the values. Used for showing
the density or intensity of data values in a matrix or table. Useful for visualizing
correlations, distribu
ions, and patterns across two
t
variables.
: Visualizing the correlation matrix between multiple variables, or temperature
Exampl
e
distribution across geographical
regions.
Heatmap
Bubble Plot
A variation of the scatter plot where each point is represented as a bubble, with the size of the
bubble indicating the magnitude of a third variable. Used to add a third dimension of data (size)
to the scatter plot. Great for visualizing relationships between three variables.
3D Data Visualization
3D data visualization is used to display data with three variables (or dimensions) and is
particularly useful when the relationship between the variables is complex, and two -d imensional
visualization doesn't provide enough clarity. With three variables, you can plot data in three -
dimensional space, allowing for deeper exploration of correlations, patterns, and trends.
The
relationship between three continuous variables by plotting data points in 3D space. Used to
explore the
interactions or relationships between three numerical variables. Example: Visualizing the
relationship between height, weight, and age of a group of people, or exploring three variables in
a machine learning dataset (e.g., temperature, humidity, and pressure).
Example: Plotting income vs. education level, with bubble size representing the population of
each group. 3D Scatter Plot
3D scatter plots can be difficult to interpret if the data is highly dense or the plot is cluttered.
Interactive plots (where users can rotate the view) are often used to make this type of
visualization more effective.
3D Line Plot
A plot where data points connected by lines are visualized in 3D space. Used for visualizing the
progression or relationship between three variables in a sequential manner, especially when the
data represents a time series or a continuous path.
Example: Tracking the path of an object in space, like the trajectory of a satellite, or visualizing a
multivariate time series over time.
Text Data Visualization
Text data visualizationtransformstextual informationinto graphical representations,
enhancing comprehension and insight extraction. Utilizing techniques like word clouds,
sentiment analysis, and network
graphs , it simplifies the data interpretation
process.
This approach turns complex textual data into visually digestible formats, supporting
effective data-driven decision-making,
relationships within the text and
data storytelling .
Text Visualization is Useful for:
Condensing a lot of content
Determining insights in qualitative data
Improved Communication
Facilitates Decision-Making
Some of the text visualization charts
• Word Cloud
• Bar chart
• Slope Chart
• Sankey Chart
• Biagram netwrok
Word Clouds
Word clouds are one of the most popular and straightforward text visualization techniques.
Display the most frequent words in a text dataset, with the size of each word reflecting its
frequency. Use Cases:
• Summarizing large text datasets.
• Identifying key themes in customer feedback or social media posts.
Eg: text = """
Data visualization is an interdisciplinary field that deals with the graphic representation of
data.
It is a particularly efficient way of communicating when the data is numerous as for
example a time series.
Excel's capabilities of managing spreadsheet d ata through data visualization tools,
such as conditional formatting and graphing tools, have made it a widely applied
data visualization tool.
"""
The larger the word, the higher its frequency. You can
use the chart (one of the text
visualization examples) to perform
exploratory textual analysis by identifying words that frequently appear in
a set of interviews, documents, or
other text.
Slope Chart
Slope charts/graphs are the perfect text visualization example when comparing time periods or
other points of reference and want to underline rises and drops across diverse categories between
two data points.
Slope Graphs (one of the text visualization examples) can be useful when you have two time
periods or points of comparison and want to show relative increases and
decreases quickly across
various categories
between two data points.
Sankey Chart
A Sankey Diagram
visualizes “a flow” from one set of values to the next. The two
items being
connected are referred to as “nodes.” The connections are labeled as “links”.
Sankey diagrams were first
used to create a visual analysis o f energy flows, but they emerged as
an effective tool for understanding the flow of time, money,
and resources. The flow design with the chart can showcase materials, fluids, revenue, energy, and costs.
Bar Charts
Bar charts can be used to visualize the frequency of specific words or phrases in a text dataset.
They provide a clear and precise comparison of word frequencies. Use Cases: • Comparing the
frequency of keywords in different documents.
• Analyzing the distribution of topics in a dataset.
• Used to show the frequency of specific terms or categories within the text.
Eg: text = "This is a sample text about data visualization. Data visualization is a powerful tool for
exploring and understanding data. It helps us see patterns and trends that might be difficult to
identify from raw numbers."
Bigram Network
A Bigram Network is a visualization technique used to illustrate the relationships between pairs
of words (bigrams) in a text dataset. This network graphically represents the most frequent pairs
of words that appear consecutively in the text, with nodes representing words and edges
representing the connections between them.
Use Cases:
Understanding the contextual relationship between words in large text datasets. Analyzing
patterns in customer feedback or social media posts to identify common themes or issues.
Exploring text data from research articles, books, or any large corpus to discover hidden
connections.