Analyzing Information Flow Through Stochastic Processes

They way information spreads is marvellous and sometimes seems almost chaotic. The word “viral” became a part of everyday language.

During MCM 2016 (Math Modeling Competition), our team analyzed how information spreads differently through information networks of society in different time periods.

2010 - mobile era.

The spread of different types of news as it originated in New York city. (red – spread, blue – not spread)

We specifically analyzed how each new era of communication technology defined the speed and qualitative change in the way information spreads.

In a gist, we used Poisson arrival processes and exponential distribution functions to model how people spread information. The key principles that we used to build our model were:

  1. Intuitively, the more people around you are talking about some news, the more likely you become to hearing about it.
  2. However, after hearing about it, you tend to lose interest in sharing that news with others as time passes.
  3. Distance makes a huge difference in the past but not so much in the present. That makes sense since the reach and speed of media sources like mobile phones and newspaper delivery are vastly different.
lambda_origin

Math Model (1,2,3 are spreading information to 4; 4 hears about it as determined by Poisson process with parameter lambda)

Read the full analysis in our paper.