Graph Analytics: A Big Data Tool for Analyzing Relationships

Graph analytics is used to study relationships between the nodes of a network. Nodes are graphically depicted as points and can represent individual people, objects, computers, or anything capable of interacting with other nodes. An interaction is represented with an arrow that connects two nodes. The arrow points from the node that initiates the interaction to its recipient. The arrows also indicate the strength of the interaction.

Strength is sometimes depicted by the arrow’s thickness. What is meant by strength depends on the nature of the interaction. If the interaction is a force between two magnets, then its strength is the magnitude of the exerted force. If the interactions are emails exchanged between a network of people, then strength in this case might be the number of emails exchanged between any two of the people.

Graph analytics is more than the mere graphical depiction of networks. The networks can be analyzed in a variety of ways through the use of graph theory. Graph analytics has a broad range of applications including biology, language, cyber security, social networks, linguistics, server networks, and search engine technology.

Nodes that have the most numerous and strongest interactions with other nodes, can be thought of as being the influencers of the network. If a plot were being hatched via email among a number of conspirators, the leaders would be indicated by the influencer nodes.

Nodes with lots of connections can also be thought of as having more exposure while those with the fewest connections are considered isolated. Great exposure could be a good thing in the case of a political election or bad in the case of exposure to disease or to other forms of risk.

Communities can be detected as clusters of nodes that have high levels of interaction. In the email conspiracy example above, communities would be the various cells of the conspirators. In networks that are incomplete, one can make reasonable deductions about possible interactions between seemingly unconnected nodes. For example, if person A and person B interact with the same people, the odds are very good that they have also interacted with each other.

Corporate Email Analysis

Businesses often keep archives of all email communication occurring among its employees as well as email communication with outsiders. Some of these businesses have employed graph analytics to analyze these communications. Graph analytics is used to construct social graphs that depict the email interactions of the employees of a company. This makes it possible to identify the company’s influencers as well as the sub networks or intra-company “communities.”

One can also study how changes of corporate email policy affect email communication. Still another use is to identify trusted email senders whose addresses are then white listed. This eliminates the possibility of their messages being incorrectly labeled as spam by email filters, thereby preventing the loss of important email.

Email social graphs can facilitate various types of investigations into fraud or corporate espionage. Tracking and analyzing email communications in this way is certainly a big data problem for many large corporations with email records that span over a decade of communication.

For more information on big data and our data storage facilities, please contact us.

Write your comment Here