Getting an Overview of a Dataset

IN-SPIRE visual tools can give you a quick overview of a dataset of documents.  Use this process to start an analysis.  

 

Documents that are close together should have similar content.  Both the Galaxy and ThemeView show the large and small concentrations of documents; both have the same spatial arrangement.

 

To get an overview of your dataset

 

  1. Open a dataset if one is not already open.  To open a dataset, from the IN-SPIRE main window File menu, select Datasets.  The Dataset Editor window will display.
     

  2. Select a dataset from the list and click Open.  The dataset will display in the Galaxy.

  3. From the IN-SPIRE main toolbar, select ThemeView.  The ThemeView visualization will display.

  4. Find major and minor themes in your document dataset by viewing the ThemeView hills and peaks.  The higher peaks mean more documents and stronger themes.

  5. The labels displayed on the ThemeView list some strong themes in that region.  To show labels, pull down the Galaxy or ThemeView View menu and select Peak Labels.

  6. Find related themes from proximity.  Peaks close together are likely related.  Isolated peaks show outliers.

  7. The Galaxy clouds and peak labels correspond to the ThemeView hills and labels.  To turn them on and off, use the Galaxy View menu.

  8. To show any strong themes in a location that is not already labeled, click on the Probe tool icon probe button and then click on a location in either the ThemeView or Galaxy.  The Probe window will display a list of theme words and the level of their contribution to the peak.

 

Related Topics