GUI Description
From SONIVIS:Wiki
Contents |
General
Analysis Perspective
In the analysis perspective there are five predefined views:
-
Characteristics
-
Statistics
-
Network
-
Node property
-
Best-of
Characteristics View
The Characteristics View displays a table with calculated domain-specific (e.g. Wiki-specific metrics, Blog-specific metrics, …) and general network metrics, organized into tabs.
By default, only activated metrics are shown there – to display all available metrics, you can use the small "Show All" button (marked by a magnifying glass) in the upper right corner of the view.
Statistics View
As statistics need some time to calculate they are switched off by default. So the appropriate views are empty when you start the tool. You can activate diagram generation in the preferences: Choose Window → Preferences... from the main menu, click on Statistics on the left and check the Activate Statistics checkbox.
If you have activated statistics various predefined diagrams are available in the Statistics View. Depending on your selection various tabs are displayed. At the moment, the following statistics are available:
| Name | Description | View | Visibility |
| Relation between authors and articles | x-axis: number of pages, y-axis: number of authors per page | Statistics View: Wiki statistics tab | Visible after database connection |
| Activity | x-axis: time, y-axis: number of pages (red), authors (blue), revisions (black) | Statistics View: Wiki statistics (temporal) tab | Visible after database connection |
| Growth | x-axis: time, y-axis: number of revisions (red), authors (black), articles (blue) | Statistics View: Wiki statistics (temporal) tab | Visible after database connection |
| Relation between author and degree | x-axis: number of authors, y-axis: degree per author | Statistics View: Collaboration tab | Only visible if network type "collaboration" is selected |
| Relation between authors and articles | x-axis: number of articles, y-axis: number of authors per article | Statistics View: Collaboration tab | Only visible if network type "collaboration" is selected |
| Relation between edits and articles of the selected node (author) | x-axis: number of articles, y-axis: number of edits | Node Properties View: Diagrams tab | Only visible if network type "collaboration" and a specific node are selected |
| Relation between edits and authors of the selected node (page) | x-axis: number of authors, y-axis: number of edits | Node Properties View: Diagrams tab | Only visible if network type "WikiLink" and a specific node are selected |
Network View
The network view visualizes network graphs. It supports three network types: Collaboration Network, Wiki-Link Network, and Dynamic Collaboration Network. Besides the network type the information space can be constrained/filtered by multiple criteria: For a MediaWiki-based network, you can choose a namespace, a Wiki category, and constrain your analysis to a given time period (via the start and end date text fields).
On dynamic collaboration networks you can also specify a number of time frames and an overlap of time frames in percent.
Node Property
The Node Property View contains a table with node metrics for the currently selected node. These properties are classified in five groups: Activity, Contribution, Miscellaneous, Page, and Position. Above these metrics, there is the node card. It shows the selected node's name and its degree.
Best-of
Based on the defined metrics in the preference page you can select node specific properties in the displayed drop-down box. After selecting a specific property this metric is calculated for the whole visualized network.
Manipulation Perspective
In the manipulation perspective there are two predefined views:
- Network
-
Graph Control
Network View
The Network View is already explained in the analysis perspective documentation.
Graph Control View
This view is organized in three parts: Display and layout settings, time frame selection, and visual filtering.
Display and layout settings
In this panel you can choose a specific network visualization layout for the visualized graph. We have implemented three different layout algorithms:
When you use force-directed layout, you can change three parameters:
- Tension
- For each edge in the network graph the algorithm simulates a spring that connects the edge's nodes. With the tension slider, you can specify a global spring tension constant that applies to all simulated springs.
- Min. Length and Edge Weight Factor
- Each spring's length is defined by
, where we is the weight of the spring's edge, a is the edge weight factor and b is a constant minimal length.
Additionally, there are different selection boxes available:
- Animation
- Switches animation of the layout algorithm on or off. Use this to stop the layout when you want to export the displayed graph as PNG, or when you want to zoom in on an interesting part of the network graph.
- Show node labels
- Activate this to display nodes as labels instead of filled circles.
- Show edge labels
- Displays the weight of each edge as label.
- Curvy edges
- This is a fun-function :-)
- Anti-aliasing
- Activate this before exporting your graph as PNG to enhance output quality. You can also activate this while animation is on, but beware, as anti-aliasing needs a lot more CPU power and decreases the visualization's frame rate.
- Highlight edge labels, Round edge labels, Show node outlines
- Controls drawing details of the visualization.
- Clustering
- If you use clustering on the graph, this checkbox controls if cluster hulls are drawn or not.
Time frame selection
The time frame selection is only available when using Dynamic Network visualization. There are two kinds of dynamic visualization: time-frame based visualization and cumulative visualisation. With time-frame based visualization only nodes which have been active in the specified time frame are drawn. When using cumulative network visualization the network's evolution is shown and the graph is continuously growing when you drag the time frame slider.
Visual filtering
The visual filtering of the network allows you to manipulate the network visually. The drawn graph can be reduced by changing:
- minimal node degree or
- minimal edge weight.
With the node navigator a specific node can be selected by its name. After selecting the node, this node is highlighted with a red border in the visualization graph.
Further Views
There are views which based on specific tool functions. At the moment there is one view:
-
Clustering
Clustering View
It is possible to cluster the graph based on two methods - partion and hierarchical.
The partition method represent a partitioning of a network into clusters by using:
The hierarchical method represent a partitioning of a network into clusters by using:
- Single linkage algorithm
- Complete linkage algorithm
- Average linkage algorithm
- Ward
Besides the definition of the applied clustering method parameters can be specified depending on the selected metrics and the number of clusters specified.
