Word Diffusion in Climate Science

Our new data mining and modelling paper is out today, “Word Diffusion in Climate Science“. Investigating the diffusion of climate science words in the Google ngrams dataset. We make observation that there is often a disjoint between the findings of science and the impact it has in the public domain. This existence of a disjoint is particularly significant when it is important the science reaches the public. Our hypothesis is that important keywords used in the climate science discourse follow “boom and bust” fashion cycles in public usage. If these cycles are linked to the science leaving the public eye then perhaps scientist need to think about they can do to ensure important findings reach as many people as possible.

Durham university press release (including a rather-too-big-for-my-liking picture of me).

Two Conference Presentations

I have been at two conferences this week. Number one was the CoSMoS workshop at the Unconventional Computation and Natural Computation conference. Where I presented a paper titled, “Going Around Again: Modelling Standing Ovations with a
Flexible Agent-based Simulation Framework”. This paper was about how I have refactored an existing simulation platform for a new but related purpose. The process that we go through to develop agent based simulations is as important as the simulations themselves. A simulation is as only as good as its worse assumption, and using the CoSMoS process can really help with identifying and documenting assumptions, and structuring the simulation in a systematic way.

Number two was at the European Conference on Complex Systems, in the Complexity in the Real World Stream. Here I presented a paper on, “Agent-based Modelling of the Emergence of the UK Banking Sector” on behave of myself and my collaborator Simon Mollan. I think a better title for this talk could perhaps have been, “A Longitudinal Study of the Population of UK Banks”. Simon collected an unique database that details the changes of a great many of the banks active in the UK. We then developed this data-set further have recently been modelling some of the population level behaviour that we have seen in the data. Slides for this talk will be made available.