A network
graph uses information from both the link and node data sets to generate
a graphical depiction of the network.
The nodes
and links in a network graph can be arranged in a variety of layout
patterns. You can choose from the following patterns:
-
Circular: the application sequentially
assigns all the connected nodes in the network within a circle.
Here is an example
of a circular layout:
-
Hierarchical: The application constructs
a tree-like depiction of the network, and places nodes with more connecting
links closer to the center.
Here is an example
of a hierarchical layout:
-
Hexagonal: The application sequentially
chooses nodes and attempts to distribute them evenly across one or
more hexagons.
Here is an example
of a hexagonal layout:
-
Multi-level force-directed: The
application combines graph partitioning and spring or force-directed
layout heuristics to determine node positions.
Here is an example
of a multi-level force layout:
Multi-Level Force-Directed Layout
-
Fixed position: The application
arranges nodes according to (x; y) coordinates that you supply with
a SAS data set.
Here is an example
of a fixed position layout:
You choose
the initial layout pattern when you create the graph. Depending on
the nature of the data, your choice of layout pattern can greatly
affect the insight you can glean from the visualization of your network.
Because you might not initially be aware of any underlying structure
in your data, it can be helpful to try multiple layout patterns to
determine which are the most useful for the network being investigated.
After creating the graph, you can later change its layout.
Once you
have created a network graph, you can analyze graph data in the following
ways:
-
Select individual observations
or groups of observations in order to focus on a subset of the data.
You can then analyze selected links, nodes, or both.
For details about selecting
data, see Selecting the Data That Appears in Graphs.
-
Use statistical graphs to subset
the data that is displayed. For example, you can create a scatter
plot based on node data and then select some of the markers in the
scatter plot. When you select markers in the scatter plot, the network
graph is updated to show only the nodes that correspond to the markers
you selected.
-
-
Regardless of the layout pattern
you are using for a graph, you can fine tune the layout by adjusting
its layout tuning algorithm. You can make this adjustment when you
set the graph's properties.
-
Use the tools available to manipulate
the graphs. For example, you can zoom in on a portion of a graph or
apply a magnifying lens to a network graph. You can also show the
labels for particular nodes in a network graph.
You might
want to change the appearance and behaviors associated with a graph
by changing the graph's properties. The properties of a network graph
determine features such as colors, line widths for links, the layout
algorithm used for the graph, and the behavior of the lens tool.
SAS/GRAPH
Network Visualization Workshop assigns colors to nodes and links based
on user-designated variables in the respective node and link data
sets. In addition, you can assign different shapes to the nodes based
on a user-designated variable in the node data set. You can assign
these variables using the Edit Data Attributes dialog box.
For details, see Specify Data Attributes.