Influence Propagation: Patterns, Model and a Case Study

<p>When a free, catchy application shows up, how quickly will people notify their friends about it? Will the enthusiasm drop exponentially with time, or oscillate? What other patterns emerge?</p> <p>Here we answer these questions using data from the Polly telephone-based application, a large influence network of 72,000 people, with about 173,000 interactions, spanning <em>500MB</em> of log data and <em>200 GB</em> of audio data.</p> <p>We report surprising patterns, the most striking of which are: (a) the Fizzle pattern, i.e., excitement about Polly shows a power-law decay over time with exponent of -1.2; (b) the<a>Rendezvous</a> pattern, that obeys a power law (we explain <a>Rendezvous</a> in the text); (c) theDispersion pattern, we find that the more a person uses Polly, the fewer friends he will use it with, but in a reciprocal fashion.</p> <p>Finally, we also propose a generator of influence networks, which generate networks that mimic our discovered patterns</p>