Predicting catastrophe with big data. Industry insight with Jack Sandford, Aon.

Jack Sanford a cyber security analyst with a background in catastrophe, explains how catastrophe modeling and big data analytics work in the real world.

Catastrophe modeling is, fundamentally, a data-driven activity: we use computational power alongside continuously refined statistical methods and professional insight from industry to assess what could happen, and what is likely to happen if disaster strikes.

Against a backdrop of climate change and population growth, insurers face growing challenges in managing accumulation risk posed by natural catastrophe. Key to addressing these challenges is having an accurate picture of where exposures are and with the advent of mega-ships and mega-ports, those accumulations are potentially moving all the time.

Accurate exposure information is not just important for predicting the future but is also fundamental to analysing the past. Without knowing what existed within the footprint of a historical event, it’s difficult to measure what the impact of that event was and therefore what the impact would be if it were to occur again.

With Skytek’s focus on data, live and historical, there is a real opportunity for insurers to develop a better understanding of a complex and ever-changing world by providing a more accurate view of exposure. Real-time vessel and port data, along with background information such as ownership and the current state of legal regimes, will allow insurers to meaningfully predict their exposure as well as the likelihood of a catastrophe.

Digital disruption is here to stay, at Aon we are embracing technology enriched with expert judgment to provide powerful insight and competitive advantage. Our ultimate aim is to drive the industry forward through innovation.

Explore how industry 5.0 is driving change in insurance. Download the exclusive Ebook. The Digital CUO Guide To Industry 5.0 – Part 1 Predicting Catastrophe. Click the banner below.

This article was first posted on