Presented by OmniSci

Conventional financial data, in many respects, is a two-dimensional view of the investment world. It provides useful facts, but without perspective. What investment professionals increasingly seek is a third dimension that captures a business in all its dynamic reality — a process, in the analytics world, that only comes when multiple forms of data are connected together, analyzed, and visualized in increasingly rapid ways.

Those forms of data in the investment world are called alternative data, or more colloquially, Alt Data.

Big Data storage technologies allow the harvesting and storage of datasets from dozens of sources — mobile devices, IoT, GPS, social media, satellite data, POS purchase data, and many others. And investors who can gain insights from these forms of Alt Data can gain an enormous advantage. The possibility of increasing speed of iteration by shaving seconds, even milliseconds, off of analytic latency accelerates the advantage.

Investors everywhere want to know more about individual company performance before others do. They are looking for leading indicators that they believe will impact future financial outcomes. Until recently, however, analysts lacked the firepower to query and visually explore these huge datasets in an effective way. Mainstream analytics software, operating on conventional CPU-based servers, simply can’t process, in a timely fashion, the billions of rows of data required for advanced, multilayer analysis.

GPUs (Graphics Processing Units), on the other hand, have finally evolved from their use in supercomputing to the enterprise data center where they can accelerate Big Data analytics, machine learning, and deep learning workloads. Designed for massive parallel querying, complex image rendering, and interactive visualization, these ultra-fast processors, working in conjunction with purpose-built analytics software, provide the kind of zero-latency speed and interactivity that professional investors require.

Investment management firms of all shapes and sizes are embracing this GPU analytics revolution. GPU-accelerated platforms are enabling speed-of-thought analysis and, in turn, uncovering more sophisticated investment and growth opportunities than ever before.

Sources of alternative datasets

The digital revolution offers many new data sources that are now being integrated for new insights. The main classes of alternative datasets are:

  • Geospatial data has location information tied into each record (e.g., address, zip code, mapping coordinates, as well as a timestamp for each event. The result is spatiotemporal data that gives analysts and data scientists the “where” and “when,” along with valuable metadata describing event details. Examples include data from mobile devices, IoT, vehicles, and location-stamped financial transactions.
  • Research/industry: Employment trends, weather forecasts, global/foreign economic data, purchasing habits, and production data from manufacturing partners can provide a broader perspective on the challenges and opportunities a company is facing.
  • Social media: The rise of social media has created an invaluable class of information, direct from customers and the public. Social networks, media sharing, consumer review apps, discussion forums, and the like can reveal important business information.
  • Additional sources: Call detail records, POS transaction data, network logs, click-through metrics on digital ads, utility smart meter logs, and similar operational data offer insight into process activity.

Once these multiple forms and sources of data are identified, the next challenge is to resolve queries from analysts at near real-time speed — a feat that is possible only through GPU-accelerated analytics. With this analytics software, analysts can examine data from scores of sources at rates impossible to achieve through conventional CPU-based systems.

Alternative Data: A new source of alpha

When non-traditional data is layered and compared, the results for decision-makers can be transformational. These insights offer new, real-time signals into how markets are moving, which in turn leads to greater confidence in investment decisions — plus the ability to outmaneuver competitors:

  • Truck destination information, shipping container locations, and credit card transactions can be used to determine net asset valuations and estimate generation.
  • Satellite views of cargo waiting on docks can help measure current and future activity for manufacturers, grocers, and transportation companies.
  • Parking lots at retail stores — another use of satellite imagery — can reveal changes in consumer purchase patterns.

In one use case, an investment research firm leverages advanced analytics to accelerate its queries into U.S. credit card transaction data. The insights gained through the firm’s unique platform are giving investment managers the ability to spot changes affecting whole industries, as well as specific companies, in near real-time — even before those changes are announced publicly.

Another example comes from an asset management firm that uses Alt Data to better understand the value of a company beyond its balance sheet. It analyzes job postings, for example, the number of data scientists and cloud computing engineers an IT company is hiring, to help determine operational efficiency and competitive edge.

Along with the opportunities, several potential risks must be noted. Data provenance is a critical factor, especially due to the wide range of sources that have recently become available. Without a thorough vetting, privacy issues, data quality, and regulatory compliance can be compromised. Investment managers need to have carefully constructed strategies to manage these inherent risks.

As the world continues to create more and new forms of data that give off powerful signals that can infer future corporate performance, GPU-accelerated analytics platforms have the potential to be a game-changing development in investment management.

Grant Halloran is an executive vice president and chief marketing officer of OmniSci

Sponsored posts are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. Content produced by our editorial team is never influenced by advertisers or sponsors in any way. For more information, contact