Amazon Web Services, Inc. (AWS), an Amazon.com, Inc. company, has announced Amazon FinSpace, a purpose-built analytics service that reduces the time it takes FSI organisations to find, prepare, and analyse financial data from months to minutes. Amazon FinSpace aggregates, catalogues, and tags data across an organisation’s data silos, making the data easily searchable by the entire organisation. The service includes a purpose-built managed Apache Spark analytics engine that contains over 100 data transformations commonly used in the capital markets industry to prepare data for analytics at a petabyte-scale.
To make it easier for FSI organisations to meet their compliance requirements, Amazon FinSpace ensures that data access controls are enforced and usage is tracked at all times. Amazon FinSpace provides an easy-to-use web application that gives analysts at hedge funds, asset management firms, insurance companies, investment banks, and other FSI organisations access to the information they need and the ability to run powerful analytics on demand across all of their data. There are no upfront costs or commitments to use Amazon FinSpace, and customers only pay for the data stored, the users enabled, and the compute used to prepare and analyse data.
Today’s FSI organisations are generating and collecting hundreds of petabytes of data every day from internal data sources like portfolio management systems, order management systems, and execution management systems— as well as third-party data feeds like high-volume historical equities pricing data, employment figures, and earnings reports. These organisations want to use the petabytes of data they possess to gain insights that help identify new sources of revenue, attract and retain customers, and reduce cost or risk.
However, before data can be analysed, FSI companies typically spend months finding the right data and getting it prepared for analysis. Discovering and preparing data is time-consuming because FSI organisations have data in silos distributed across departments that specialise in particular assets or geographies and generate specialised data (e.g. equities, options, bonds, European mutual funds, Asian currencies, etc.). Furthermore, data access is tightly controlled by regulation and policy, meaning analysts must justify to compliance officers how their access will conform to data use policies before they can access the data. Once they are granted access to the data, analysts must prepare it for analysis by iteratively performing data transformations to discover new insights within the data.
For example, capital markets traders often use technical indicators like Bollinger Bands, Exponential Moving Averages, and Average True Range to identify undiscovered trends and patterns. Many of the data analytics tools available to analysts today were built to run on a single computer and were not designed to take advantage of the cloud’s scale and the ability to compute heavy analysis on-demand. As a result, analysts either have to use small representative datasets that limit predictive ability, or the data has to be manually broken up into many subsets, transformed piecemeal, and manually recombined. Neither approach is ideal or effective.
Amazon FinSpace solves the challenges FSI organisations face by vastly simplifying the steps needed to find, prepare, and analyse data, reducing the time involved from months to minutes. Customers begin by ingesting data into Amazon FinSpace from internal data silos or third-party data feeds via the service’s Application Programming Interface (API) or a drag-and-drop interface in the web application. To find data, customers simply browse a visual catalogue and search for familiar business terms like options trades for the last three years or US automotive bonds from within the web application.
Amazon FinSpace includes built-in classification schemas for common FSI data sources (e.g. trades, corporate actions, and economic data) that customers can customise to their needs, so the data can be organised in a way that is easy to find and share. Amazon FinSpace records the daily updates and corrections received for datasets and processes them to create point-in-time views to validate modelling assumptions and to show what data was used to inform past decisions for historical analysis. Customers can use built-in Jupyter Notebooks to access data stored in Amazon FinSpace and can then choose from over 100 built-in functions to prepare their data for analysis (e.g. Bollinger Bands, Exponential Moving Averages, and Average True Range)—or they can build and use their own functions to prepare data for analysis.
Amazon FinSpace provides managed Spark clusters that can be scaled up or down on demand so organisations can benefit from the elasticity, scale, and cost savings provided by cloud computing. Customers define their data access policies within Amazon FinSpace, and the policies are automatically enforced across data search, visualisation, and analysis. Amazon FinSpace records data access, tracks data usage, and generates compliance and activity reports indicating who accessed data at what point in time.
“FSI organisations generate and purchase massive amounts of data, but using this data is very difficult because of the time and effort it takes to collect and prepare data for analysis,” said Saman Michael Far, VP of Financial Services Technology, AWS. “Amazon FinSpace is a game-changer for FSI organisations. Amazon FinSpace radically reduces the time it takes for FSI customers to do analytics across petabytes of data, making it significantly easier for them to identify new sources of revenue, attract customers, and reduce cost and risk.”
Amazon FinSpace is generally available in US East (N. Virginia), US East (Ohio), US West (Oregon), Canada (Central), and Europe (Ireland), with plans to expand into additional regions.