Arrange Networks

and Trees
 Networks?
Network data?
Networks
• Network visualization, also referred to as graph visualization or link
analysis, is a method of visually representing relationships between
various elements or entities within a network.

• This type of visualization simplifies the complex nature of the network,

making it easier to understand, analyze, and interpret: essentially, a
picture is worth a thousand words.

• Nodes: The individual elements or entities within the network. In a

social network, for example, nodes could represent people.

• Edges: The connections or relationships between nodes. These could

be one-to-one connections (pairs) or more complex associations.
Types of Networks: Networks can be classified based on the relationships:

Social Networks: Represent relationships or interactions between people.

Communication Networks: Represent pathways for data transmission, such as the

internet.

Biological Networks: Represent relationships between biological entities, such as gene

interactions.

Topology: The structure of a network, which can be:Star, Ring, Mesh, etc. Each
structure influences how information flows through the network.

Weighted Networks: In these networks, edges have weights, which could represent the
strength, cost, or frequency of the relationship between nodes.
Network data
Data that describes the structure, relationships, and attributes of nodes and edges within
a network. It's often used to analyze interactions, flow, and connectivity patterns.

Nodes Data: Information about the entities in the network (e.g., name, attributes, or
categories).Example: In a social network, node data could include a person’s name, age,
and location.

Edges Data: Information about the relationships or connections between

nodes.Example: In a social network, edges may represent friendship, and edge data
might include when the relationship was formed or the strength of the connection.

Degree: The number of connections a node has. A high-degree node has many direct
connections.

Adjacency Matrix: A common way to represent network data, where rows and columns
represent nodes
Why is network visualization important?
Network visualization comes with several advantages that can be a game-changer for
across industries.

Gain insights faster

The human brain processes visual information much faster than textual information.
Visualizing a network makes it much faster to understand key information around
relationships within a dataset, hierarchies of information, etc.

Intuitive understanding of your data

Exploring connected data within a visualization is much more intuitive than trying to
sort through spreadsheets or tables.visualizing is a fast and powerful way to analyze
your data.
Flexible and dynamic
A network visualization tool can give you a dynamic way to visualize, explore, and
understand your data. This flexibility enables you to gain key insights faster and more
effortlessly.

Easily sharable insights

The saying “a picture is worth a thousand words” is certainly true when it comes to
network visualization. An image is a more impactful way to share key findings and
information with decision-makers and other stakeholders.
Networks and
Graphs
Applications
● Social network
analysis

● Biological
applications
https://www.wien.gv.at/stadtentwicklung/projekte/verkehrsplanung/u-
bahn/planung sergebnis/ [January 6th, 2020]
●
● Software
Transportation
engineering
networks

● Communication
networks

● Security networks https://www.researchgate.net/figure/Social-Network-Graph-of-Festival- 6

Host-Organiz ations-and-Partners_fig1_313389833 [January 6th, 2020]
Example
Example
Example

Confusing Much more

representation clear!
Topology Tasks: Examples

Finding
○ all possible paths from one node to another

○ the shortest path between two nodes

○ the adjacent nodes one hop away from a target node

○ all nodes that act as a bridge between two components that would otherwise
be disconnected
Example: Mobile phone network

Nodes: people
Edges: phone
calls
Example: Transport network

Nodes: places
Edges: train
lines
Example: Transport network

Nodes: places
Edges: train
lines

Shortest
path?
Example: Historical
data
Wher Wher Whe
e e n
origin origin signing
current place date
city mentioned Docume creation
Perso nt date
n

Wha Who Wha

t t
author
professio contract contrac
n gender or t tweet
organizati witness book
on certifica
te
Contracts: who signed a contract with
whom?
represent it as a
graph

link (edge) between 2

nodes

nodes
(vertices)
Contracts: who signed a contract with
whom?
Node-link
diagrams
Contracts: who signed a contract
with whom?
● Node-link ● Matri ● Biofabr
diagrams x ic

Node-link layout Tabular layout

2
4
Multivariate Networks

• gender: female • gender: female

• profession: • profession:
shopkeeper designer
• hometown: London • hometown: Paris People have attributes -
• age: 42 • age: 33 how to visualize them?

• gender: male • gender: male

• profession: • profession:
shopkeeper photographer
• hometown: Rome • hometown: Madrid
• age: 47 • age: 36

2
5
Mapping people attributes to visual
attributes of nodes
● Color ● Node
intensity size

• age: • age:
43 30

• age: • age:
50 37
 Mapping people attributes to visual
attributes of nodes (On-node encoding)
● Color ●
Symbols

• gender: • gender:
female female

• gender: • gender:
male male
gende gende
r r
femal femal
e e
male male

Nobre, C., et al. "The state of the art in visualizing multivariate networks." Computer Graphics Forum. Vol. 38.
 Mapping people attributes
● Attribute-driven
faceting(Commo
● Contour
n Environment)
overlay(Grouping or
structure) femal mal
e e

• gender: • gender:
female female

• gender: • gender:
male male
gende
r
femal
e
male
 Mapping people attributes

● Geographical attributes (Attribute-driven

positioning)

• hometown: • hometown:
London Paris

• hometown: • hometown:
Rome Madrid
Contracts have attributes

• payment:
140
• type:
service

• payment:
80
• type:
goods
• payment:
130
• type:
service • payment:
90
• type:
goods

3
0
Contracts: how to visualize contract's
attributes?
● Edge
attributes

type
servic
e
good
s
paymen typ
ts e
3
1
Comparison of Node-Link Layouts

Modifying visual Grouping nodes Assigning node or

appearance of a node according to one or edge positions
Summary:
or an edge more attributes and according to one or
or embedding marks in places the elements more attribute values
a node of a group in a shared
region
● Free layout (e.g. ● Relaxed ● Fixed layout
force-directed or orthogonal) fixed
Layout:
● Styled layout (position layout 3
based on pattern, e.g.
Nobre, C., et al. "The state of the art in visualizing multivariate 2
Connection vs. Containment

• Relevant for drawing trees

• Connection -- traditional view of graphs
• Containment -- essentially treemaps
• Type of information is different:
– Connection marks: Only show pairwise relationships between
two items
– Containment marks: Show complete information about
hierarchical structure
3
3
 Connection vs. Containment
Dataset: 5161-node tree (computer file
system)

Rectangular horizontal node-link BubbleTree node-link Treemap showing hierarchical

layout (connection) layout structure with containment
(connection) rather than connection
3
4
Connection vs. Containment
Eight visual encodings of the same
tree dataset

3
5
Connection vs. Containment
Compound network: Combination of a network and tree

Example: GrouseFlocks (uses containment to show graph

hierarchy structure)

(a) Original (b) Cluster hierarchy built atop (c) Network encoded using
graph the graph, shown with a node–link connection, with hierarchy encoded
layout. using containment.

3
6
Limitation
Scalability Issues: size of the network grows (i.e., the number of nodes and edges
increases), the visualization can become cluttered, making it hard to interpret.

Interpretation Complexity: Large, intricate networks may be difficult to interpret, especially

if they lack clear structure or context.

Edge Crossing and Clutter: In large networks, it can be difficult to distinguish between
edges if they overlap or cross each other.

Lack of Standardization:There is no universal agreement on how certain types of networks

should be visualized, which can lead to inconsistencies

Over-Simplification of Data: While simplifying complex data can make it easier to interpret,
it can also result in important details being lost in the process.