Social Network Analysis

Challenges

 

Nothing is completely isolated. Any element interacts with the environment or with other elements. These relationships provide information about the social behaviour of both customers and the network of business agents in the environment, whose intelligent exploitation generates significant benefits in very different applications.


Approach

 

This science provides a wide range of metrics on graphs that detect different characteristics of the social network’s own interpretation, in the context of a social network, without the need for a theoretical framework.
Social Network Analysis is the subject of study within computer science because the algorithms are implemented in practical cases of networks of hundreds of thousands of nodes, or in some cases millions of nodes.
Nothing is completely isolated. Any element interacts with the environment or with other elements. These relationships provide information about the social behaviour of both customers and the network of business agents in the environment, whose intelligent exploitation generates significant benefits in very different applications.
Two of the main objectives of the SNA are the detection of influential individuals in a social network and the detection of associations of individuals by common interests (communities).
The natural flow of the use of these metrics is usually as follows:

  • Data is obtained from a social network
  • The social network is read into as a graph
  • SNA metrics analyse the graph
  • SNA results give us new information about the original social network.

 


In the following example, taken from a talk by Jure Leskovec at Stanford University we can see a graph created from the facebook social relations of one of his students.

In colours, we see the groups that have been detected using a clustering algorithm,
specifically the AGM (Affiliation Graph Model).

In black colour the cases that match the detected cluster coincides with a particular real life social group, in white colour, incorrect cases.

Although the algorithm only knew the relationship information, the groups that have
been found have a perfect match with the different social environment, which involved the initial user: School, Work, Squash and Basketball.



Benefits

 

Using Social Network Analysis enables the generation of high-value content like:

  • Generation of Public Objectives for Viral Marketing Campaigns
  • Prevention churn
  • Content recommendation systems
  • Estimated evolution of social trends
  • Interpretation of social relationships

 


Success Cases

 

  • Persons System of social network analysis in Telecommunications operators
  • Churn prevention System in telecommunications for Spanish operators
  • Determination of public objectives for viral campaigns in a virtual operator in Spain
  • Analysis of the correlation between social network position and academic performance for a distance university.