Storygraph Visualizer

The Storygraph Visualization Tool (Storygrapher for short) presents a set of stories in graph format. The current build uses hard-coded input files to construct its graph, which is based on the stories seen in 'summarized results.' If you download the stand-alone application, be sure to read the 'readme.txt' before running it.

You can interpret the graph using the following key:

•   PubMed Document
•   PubMed Document (Beginning of story)
•   PubMed Document (End of story)
•   Frequent Episode (Essentially a cluster of documents)
•   Story Edge (Connects two adjacent documents in a story)
•   Thick Story Edge (The thicker the edge, the more often it occurs in the visible stories)

Storygrapher contains three main panels, which we describe individually:

• The graph panel: displays nodes and edges representing the visible user-selected stories
  
  
  
  
  To interact with the graph you can click-and-drag a node to move it, single-click it to display its information in the Info Panel (see below) or double-click it to highlight its containing stories. Clicking on a Frequent Episode node (box) will display the documents it contains in the Frequent Episode View (see Control Panel below). Note that selected nodes are highlighted in yellow.
• The control panel: Allows graph layout selection and story visibility management, and contains the Frequent Episode View
  
  
  
  
  The drop-down menu at the top of this panel allows you to select different graph layouts. Some of these are just for variety, but we strongly recommend using Story Layout. This is a modified version of the Spring Layout that automatically arranges stories from left to right (start to end). After 30 seconds the graph will stop moving, allowing you to arrange nodes how you see fit.
  
  The table contains a list of the available stories, each of which can be made visible by clicking its check-box. Note that the "Story #" corresponds to the story numbers seen on the 'summarized results' and 'full top results' pages. You can show and hide stories as described by the four buttons below the table, and a brief explanation is provided by clicking the '?' button.
  
  The small panel below these buttons is initally blank but displays the nodes in any Frequent Episode (box node) that you select. The node annotated with an asterisk (*) is always the leftmost node in the Frequent Episode. These Frequent Episodes are essentially frequently occurring sequencies of documents that we have condensed.
• The info panel: Displays basic information about the selected node or frequent episode
  
  
  
  
  This is the simplest of Storygrapher's panels, displaying information about the node you have currently selected. For a document this panel displays its PubMed ID (which is an active link to the article in your default browser), title, authors, and keyword(s). These keywords are automatically assigned based on LingPipe's named-entity recognition engine and the MeSH headings for each document.
  
  For a Frequent Episode this panel simply displays its ID.