We’re excited to share our latest customer case study, about how KatRisk, a leading catastrophe risk modeling firm, used Domino to deploy its models more rapidly to its large reinsurance clients.
At KatRisk, our focus is on delivering the most accurate flood and wind risk models possible. Domino helped us get our products to market faster, and provided a full platform for our data science needs.
Dag Lohman, CEO of KatRisk
KatRisk provides comprehensive and cost effective catastrophe risk models to clients. Their models have an open architecture allowing users to better understand underlying model parameterizations and even modify models to suit specific needs. Current products cover flood and tropical cyclone risk in the US, the Caribbean, and Asia.
Best-in-class models, a large customer, and a challenge
KatRisk had developed industry-leading models, written in R, for predicting flood and wind risk. These models were innovative in their accuracy and in the way that clients could customize them for their specific needs. KatRisk also had a customer: one of the world’s largest reinsurance companies wanted to use the models in automated processes for pricing reinsurance policies.
KatRisk’s models were sophisticated, constantly improving, and depended on nearly 1TB of data. The reinsurance company had no easy way to handle delivery of these models. Because the models would be invoked programmatically in automated processes, they needed to be highly reliable, even as KatRisk released updates as their models evolved.
The team at KatRisk wanted to deliver products as quickly as possible, and needed a stable, mature platform to deliver those models where clients could access them programmatically.
Looking at their requirements, the team at KatRisk recognized that building the functionality was technically possible, but would take time, and distract them for their core mission. Instead, they chose Domino’s data science platform.
Domino’s “API Endpoint” functionality let KatRisk take their existing R code and, without changing it at all, deploy it as a production-grade REST API. Domino’s platform handles high-availability, SSL, zero-downtime upgrades, and more. KatRisk’s reinsurance clients can then invoke the KatRisk models by calling a simple URL, no technical integration required.
The Domino web interface lets you publish your R models with one click, without any custom packages or changes to your code. Once published, Domino provides code snippets in a variety of languages showing how to invoke your model.
Getting their existing models working in Domino was extremely easy. Because Domino is an open platform, they didn’t need to rewrite any code. They could also continue to use the development tools they were already using or change their workflows.
As a result, within days, KatRisk had their models available for consumption by their reinsurance client.
Results for KatRisk
After a matter of days, KatRisk’s models were integrated into real-time policy pricing algorithms at a major reinsurance company. By using Domino to accomplish this, KatRisk saved months of development time, and many tens of thousands of dollars in engineering and operational costs. They got all the functionality they needed and more. Most importantly, by choosing Domino as their platform, they could focus their resources on creating the next generation of risk models instead of building infrastructure.
Want to see Domino for yourself?
Register for a live demo to see Domino in action and ask questions.
New to Domino? Consider a Guided Tour.Start Your Free Domino Trial
Recent PostsTensorflow, PyTorch or Keras for Deep Learning Reinforcement Learning: The K-armed bandit problem KNN with Examples in Python Computer Vision in Deep Learning: An Introductory Guide Supervised vs. Unsupervised Learning: What’s the Difference? Powering Up Machine Learning with GPUs A Guide to Natural Language Processing for Text and Speech What Is Reinforcement Learning and How Is It Used? A Detailed Guide To Transfer Learning and How It Works Introduction to Deep Learning and Neural Networks
Other posts you might be interested in
Subscribe to the Data Science Blog
Receive data science tips and tutorials from leading Data Scientists right to your inbox.