Lecture 9 Flashcards
Relationships Among Data
Sometimes, while working on data we find that a relationship exists between different data elements.
These relationships can be hierarchy, network etc.
To represent these kinds of relationships, we use tree maps, heatmaps, dendrograms, network diagrams etc.
These relationships can be simple or complex, unidirectional or bi-directional, weighted or non-weighted, certain or uncertain
Displaying Hierarchical Structure
Trees or hierarchies are one of the most common structures to hold relational information.
For this reason, many visualization techniques have been developed for display of such information.
These visualization techniques can be broadly divided into two categories:
Space-Filling Methods
Non-Space-Filing Methods
Space Filling Method
As the name implies, it makes maximal use of the display space.
This is accomplished by using juxta-positioning to imply relations, as opposed to conveying relations with edges joining data objects.
The two most common approaches to generate space-filling hierarchies are :
Rectangular Layout
Radial Layout
Rectangular Layout
The most popular space-filling rectangular layout is the treemap.
These are the alternative representation of the Venn diagrams.
In the basic treemap, a rectangle is recursively divided into slices, alternating horizontal and vertical slicing, based on the populations of the subtrees at a given level.
Many variants of treemaps are available today like:
Squarified treemaps – to reduce the occurrence of long, thin rectangles
Nested treemaps – to emphasize the hierarchical structure
Radial Layout
Radial space-filling hierarchy visualization have the root of the hierarchy in the center of the display and use nested rings to convey the layers of the hierarchy.
Each ring is divided based on the number of nodes at that level.
The number of terminal node in the subtree determine the amount of screen space that will be allocated to it.
However, unlike treemaps, which assign most screen space to convey terminal nodes ,radial techniques also show intermediate node.
Non Space Filling Methods
Most common way to represent tree is a node-link diagram.
Have two factors:
Fan-out degree – Number of siblings a parent node can have.
Depth – the furthest node from the root
Network Diagram
Network diagrams are made up of two components: Node and Line.
Node is also known as vertex, site or actor of the diagram.
Link is known as edge, bond or tie of the network diagram.
Network Diagram Properties
The number of immediate connections of a node provides the degree property of that node.
In the case of directed networks, degrees are designated as
“in degree”—the number of links destinedtothe node, and
“out degree”—the number of connections originated at the node, orfromit.
Path and Connectivity
In order to understand distances in a network, scientists developed the concept of a path that is any sequence of nodes given that each consecutive pair of nodes is connected by a link.
The path length provides the number of links in the route between a pair of nodes.
When there are no paths between a pair of nodes, it means that a network is not connected and it is divided into subgroups, called “components” in network science.
Other metrics were devised for questions related to distances; the shortest path between two nodes and the network diameter are two examples.
Network Layouts
The common types of network layouts are: Linear Force Directed Circular Linear Matrix Radial Community Structure
Linear
Nodes are organized linearly and the links are usually arcs connecting nodes.
Con:It’s hard to identify clusters and is only feasible for small datasets.
Force Directed
There are many algorithms that use an iterative process to locate nodes according to physical forces.
They are centred on a single node.
Con:There are too many node occlusions and link crossings in dense areas.
Circular
Nodes are organized around the circumference and usually grouped by categories. Links cross the circle and are usually bundled so as to simplify the crossings.
Con:It’s hard to identify clusters.
Polar or Radial
Nodes are organized around a central node, with their position related to the number of hops it takes to reach it.
