Visualizer
From Piki
Visualizers are tools used to visualize data in Synapse. They are used in preprocessing mode.
Contents |
Operation
A visualizer take a data unit and displays its contents in some fashion. It has capabilities to handle selections and can deploy filters to the filter stack. It can also pass along a filtered version of the incoming data unit along with a selection to be used by another visualizer. This enables Synapse's hierarchical visualization.
Usage
To select a visualizer you must first be in preprocessing mode and either have a data unit to visualize, or another visualizer that you want to link.
Visualize a data unit
Start by making sure you have an empty visualizer tab open. Then drag the data unit from the data unit manager and drop it on the empty visualizer tab.
 |  |
As you drop the data unit a pop-up selection window will appear. Click on the visualizer you want to use.
Link to another visualizer
This procedure only differs slightly from the one above. Instead of dragging the data unit from the data unit manager you drag the tab of the source visualizer to the empty visualizer tab. For this to work, you must grab the actual tab, the part that displays the name of the source visualizer.
After two visualizers are linked, what is selected in the first will be reflected in the second. This hierarchical visualization is extremely powerful for data mining purposes.
Common features
Since a data unit has a training and a validation set visualizers usually have a 'set' drop down selector in the visualizer menu bar. Many visualizers also have the ability to save the data displayed. Either as CSV text file or as an image.
As visualizers are seen as a transient dynamic tool their state is not saved when the solution is saved.
See also
- Preprocessing - An article on the Synapse preprocessing mode.
- List of visualizers - A list of available visualizers in Synapse.