Relationships - Graph Analysis and Visualization - page 337

Almost all of the graph is tied up in one giant component (top) with too
many connections to reveal any patterns. Other teeny clusters indicate films
with connections between the co-stars but otherwise not connected to any
other film. For example, the large yellow dot near the bottom center
represents Adolf Hitler, which is connected to other Nazi generals via links
that represent the movie Triumph of the Will. These actors did not appear
in any other films, which is why there are no connections to other movies.
Another large yellow dot is labeled “Napoleon,” which seems surprising
because Napoleon died before the era of moving pictures. In this case,
“Napoleon” refers to an Indian actor named Kumaresan Duraisamy, whose
stage name is Napoleon. This little example hints at some of the challenges
when using open source Big Data. Though it is promising, you are prone to
experiencing a wide variety of errors, omissions, and anomalies when using
open source data such as Wikipedia, as opposed to a better curated data set.
Extracting a smaller network for a specific analysis may be closer to a
specific objective, such as an analysis of the co-stars of Leonardo DiCaprio.
Figure 9-11 shows the ego network corresponding to Leonardo DiCaprio.
You can extract an ego network in Cytoscape by following these steps:
1. Select the ego node by entering the node name in the search box.
2. Expand the selection using Select First Neighbors of Selected Node,
available as a button on the toolbar or under the Select ⇒ Nodes menu.
3. Put the selection set into a new graph window using New Network from
Selection, available as a button on the toolbar or under the File ⇒ New ⇒
Network menu.
4. Apply a layout algorithm to this subgraph from the Layout menu.
Each node represents an actor. Each link represents a movie in which two
actors starred. Straight links indicate a single movie. Curved links indicate a
second or third movie in which both actors co-starred. (You can mouse over
any link to see the name of the movie.) The nodes are sized by the number
of links the actor's Wikipedia page has, and color is based on the number of
times the actor appears in the overall data set. Thick lines indicate movies
that have more links in Wikipedia (for example, The Aviator or Inception ),
as opposed to movies that have few links (for example, Don's Plum ).