From Dickens to Data Science - Making the Case for Transcendental Information Cascades

2019-12-31T20:42:59Z (GMT) by Markus Luczak-Roesch

In this talk I introduce Transcendental Information Cascades, a method to understand the temporal dynamics of naturally occurring complex systems through the lens of a specific kind of spatio-temporal network that represents information token recurrence. The method suits very well for exploratory data analysis when it is unknown what is going on at the very low levels of a system's dynamics, and allows switching between looking at individual events or accumulated event sequences. I will present one application of this method in a digital humanities project where we analysed the entire corpus of Charles Dickens's novels and discuss the potential it has for other application areas such as the analysis of brain wave recordings, patent mining or the study of online communities for example.