Economic Sentiment Indicator Algorithms: Predicting Brexit-like Events Using Big Data

Project Description

Zurich Insurance is a global insurance company. Our Investment Management team has several challenges where Big Data and the underlying technology can play a role. For instance:

• Outlooks on macroeconomics and financial markets are currently developed based on market indices, market economic indicators and a selection of research papers. These outlooks are also incorporated in the tactical asset allocation for our own assets and therefore can potentially generate great financial benefits.

• We believe we can complement current information with additional “non-traditional” information on economic and financial developments consisting of both structured data (e.g. time series, research pools) as well as unstructured data such as social media, Google search requests, geo-data, etc. However the vast amount of data is simply too much for traditional methods.

• Brexit is a good example where we could back-test a possible solution. With Brexit all "white-collar" reports and research indicated Brexit would not occur. We therefore would like to test a solution that could ingest and process large amounts of unstructured data and determine an economic sentiment indicator. In this case applying cluster analysis to the data to find economic turning points pre-Brexit.

We would like to develop a big data system with a set of algorithms that would give us predictive capabilities on a macro-economic level.  We can share more details with the experts with the best proposals at a later stage in the hiring process.