How Azure Helped Munich Re Pioneer Predictive Modelling

Munich Re, a global reinsurance firm, depends on analytical models to assess risk. It has been calculating weather-related risk, using a huge number of parameters for over a century. The data it gathers gives the company a close-up view of natural disasters and their potential to cause physical and financial harm. The problem soon arose that the company had too much data and not enough time or computing power. Fortunately, the advent of cloud computing changed all of that, giving the company a way to finally analyse huge volumes of data in a comprehensive way. With the help of Microsoft Axure, Munich Re has been able to engage data scientists and engineers to examine its data and find new approaches to risk assessment and climate research. 

What Challenges Was Munich Re Facing?

Munich Re is german reinsurance company, offering insurance to the insurers who can become overwhelmed with hefty payouts caused by natural disasters and global weather events. As one of the world’s leading reinsurers, Munich Re has been collecting data around natural disasters for centuries. With such a close eye on natural phenomenon, the company soon started noticing  rising temperatures, melting glaciers and other signs of climate change. So, since the 1970s, more data was collected in an effort to understand the causes of global warming, and ways to reverse them. 

Way back in 1973 the company wrote a publication of floods stating that climate change could alter the risks of natural disasters. The company has now been dealing with climate change for over 40 years, trying to understand the drivers in order to find solutions. Its database includes records of damage caused by earthquakes, volcanic eruptions, heat waves and more and the company even has its own corporate climate centre. 

The company’s experts have been developing statistical models for over 20 years with the aim of simulating potential disasters and creating possibility curves. Data is key to Munich Re’s predictive models, but it needs to be analysed, assessed and ultimately understood to provide value. However, the problem is that with that much data the queries set up for analysis took several hours, and the supporting hardware was extremely expensive. While over time the computing power in-house increased, largely due to virtualisation technologies, data processing still took hours. And, the amount of data continued to grow with additional raw data from the likes of NASA satellites added to the mix. The processing of over 3,000 terabytes of data from 40,000 events required a powerful infrastructure that Munich Re couldn’t deliver in-house. It needed to find a new way to analyse its data and to keep up with predictive modelling; the cloud was the obvious next step. 

Moving Data Analysis to the Cloud with Microsoft Azure

The cloud was a huge breakthrough for Munich Re, providing the tremendous computing power that the company needed. Microsoft Azure has been powering the company’s predictive modelling since 2016 with flexible storage solutions and virtual machines that let them increase computing power virtually and on-demand. The company doesn’t have to invest in CPU power until required, making it much more affordable. 

In-memory processing has been a game-changer for Munich Re, reducing several hours of scenario calculations down to a few short seconds. The processing is exponentially faster, and that means the team can try out more models in the same amount of time. What’s more, the collaborative nature of the cloud environment means that scientists and business analysts can work together with the same data at the same time and create interdisciplinary processes and improved results. 

The company continues to generate data, and this will only continue into the future. Fortunately, with Microsoft Azure, Munich Re has a clear path for economically saving and processing data, and this creates further opportunities. The company’s data analytics team processes thousands of times the data than it did a few years ago, combining raw data from satellites with unstructured data from word and image files. They are also starting their first projects in the area of automatic video analysis. 

The Benefits of Partnering with Innovation

Azure is an innovative platform that delivers powerful remote workstations, big data analytics and machine learning to organisations like Munich Re. It’s AI capabilities help to accelerate scientific discovery with the intelligent cloud and intelligent edge. With the power delivered by Azure, companies are able to perform cutting-edge research with intelligent capabilities, including:

• Deep learning – running compute-intensive deep learning workplaces can deliver near-infinite CPUs and seamlessly create simulations in the cloud.
• Advanced analytics – machine learning models allow more insights to be derived with inferencing models on top of simulation data.
• AI research – with Microsoft at the forefront of creating next-generation intelligent numerical models, scientists have unprecedented access to tools and infrastructure. 

Could Microsoft Azure Help Your Business?

Microsoft Azure has helped Munich Re revolutionise the way it looks at data. Just a few short years ago, what they are achieving today would have been impossible. The cloud has changed everything, allowing them to analyse risks on a new level. And, Microsoft Azure provides a level of innovation that meets their expectations while at the same time giving them a cost-effective way to manage huge volumes of data. In today’s digital world, every business collects data. However, it only has value if you can store, organise and analyse it. Microsoft Azure can help you do just that, unleashing insights that could change your business, open up new opportunities and give you the competitive edge you’ve been looking for. 

To summarise, here are the challenges, the solution and the results that Munich Re experienced:

• Challenges – Munich Re is a global reinsurance firm that relies on data analysis to assess risk. It uses a vast number of parameters and has been gathering more and more data to input into its predictive models. However, data analysis took several hours, and the supporting hardware was very expensive. The company needed to find a way of processing over 3,000 terabytes of data with a powerful infrastructure.
• Solution – Microsoft Azure proved to be the perfect cloud partner for Munich Re. It offered flexible storage and virtual machines to exponentially increase computing power. As the company didn’t have to invest in CPU until it was needed, analysis became much more affordable. What’s more, processing is incredibly fast, and the cloud-based nature of data storage encourages interdisciplinary collaboration. 
• Results – the advent of the cloud was a game-changer for Munich Re, giving it the computing power it required to continue to create innovative predictive models. Azure has given the company a clear path to continue collecting, storing and analysing data in an economical way. The team now processes thousands of times the data that they did a year ago and are just beginning their first projects in automatic video analysis, an impossible task without the power of the cloud.