Implementing Real-Time Data into City Risk Index
🚄

Implementing Real-Time Data into City Risk Index

🗒️ Background

Lloyd's of London is an insurance and reinsurance market located in London, England. They develop market analytics for all insurance brokers that do business in the United Kingdom. This consulting project was led by our Managing Partner Saif Bhatti.

image

🛠️ Challenge

The Lloyd’s City Risk Index shows how much economic output 279 cities would lose annually on average (GDP@Risk) from 22 man-made and natural threats. They collaborated with the University of Cambridge to construct a dataset with 10-year projects for risks from various risks, including adverse weather events, political instability, and market shocks.

This tool featured two views:

  1. External View: Market analytics for the greater market
  2. Internal View: Connecting City Risk Index with proprietary Lloyd's data one internal views that makes use of proprietary Lloyd's data to allow underwriters access to cutting-edge analytics.

The dataset developed features 10-year predictions, but this data is static and is continually going out of date. In order to keep the Internal View dynamic and useful for underwriters in real-time, new data was needed to update the underlying risk models.

💡 Solution

The solution involved developing live risk monitoring by incorporating API feeds for the existing man-made and natural threats the risk models account for. In sourcing this data, he turned to APIs USGS, NOAA, as well as web-scraping news sites like Independent and Times.

He developed automated data collection and computation upon capture. If this data came from a structured API, it was melded into the data lake. If the data was scraped or extracted in an unstructured manner, then it was first synthesised using NLP methods to determine whether it held specific information that could geographically associate it with a city (within the relevant set of 279 cities), and then tag it with a threat.

These feeds served as live risk monitoring, to help underwriters make use of projection data, as well as incorporate tail-end event.

He also aided Lloyd's in vetting new data vendors and engaging with underwriting teams to gauge use cases and efficacy.

📊Results

The developed data flow successfully adds live risk monitoring using Python to interact with the relevant APIs and web scraping tasks, ultimately adding to a data lake visualized with Qlik Sense. He contributed to Lloyd’s of London’s City Risk Index v2.0, open for public release.

Let's Connect

Interested in collaborating? Get in touch

©️ Go Fish Analytics 2021 - All Rights Reserved - Built with ❤️ in Austin, TX