Information Analytics

One of the biggest opportunities in business today is transforming the vast and growing universe of enterprise information into immediate, relevant insight. We all know the amount of digital information exploding, and at the same time expectations for processing this information are rising. Organizations rely on business analytics to make decisions about everything from internal operations to customers, sales, and supply chains—essentially anywhere that actionable business intelligence can cut costs and improve efficiencies.

Valuable content is distributed across the enterprise, 95% of it in “unstructured” formats—that is, residing not in databases but in a variety of files that include email, documents, presentations, and much more. To have a complete view of the business, companies need analytics that leverage both structured and unstructured information.

At HP Labs, we are working to redefine the twin tasks of taming and exploiting this information to revolutionize enterprise decision making. We’re applying mathematic and scientific methodologies and developing the models, tools, and algorithms that drive informed, highly effective, repeatable, and consistent decision making---to ultimately create better-run businesses. Our research enables near real-time business intelligence with robust, scalable data management, data-intensive analytics of structured and unstructured information, and automatically delivering information described in the language of business. 

Information Analytics Big Bets:

Context Analytics

Context-aware computing leverages information about an end user or system – information like environment, processes, communities, roles or identity – to provide personalized interactions that ease discovery of content and services (or processes, in the case of systems) and dynamically adapt as users change context. This capability will be key to achieving user satisfaction or improved system performance in future IT services. Existing context-aware applications typically provide application-specific solutions, focusing on specific contexts and information processing needs, and often consider only simple contexts such as location and time. Our research studies a wide variety of context types, particularly sophisticated contexts (beyond location and time), and develops generic programming and systems support for context-aware computing, allowing for shorter development cycles with higher quality. Applications may include solutions for consumers (personalized printing), enterprise (business processes), and systems (networking).

Enterprise Swarm

The Enterprise Swarm is a productivity platform that creates value from big data in enterprises, and delivers this value to users by streamlining collaboration, creating insight and increasing automation. Our research aims to tap into big data in the organization, aggregating unstructured and structured data including knowledge bases, communication streams, transactional data, performance monitors, and user behavior, among others. Behind it all are powerful analytics that are domain-aware and context-dependent, and that yield personalized information.

Live Analytics

The Live Analytics project aims to deliver predictive analytics and high-quality insights that transform operational business processes via data-flow and streaming analytics over structured and unstructured data. Our research explores several critical analytics applications that may be delivered as business solutions, such as our work on Live Customer Intelligence, Live Operational Intelligence, Healthcare Analytics, and Enterprise Document Content Analytics. Finally, we are designing novel approaches to visual analytics, providing users with richer, more useful visualizations of their data.

Taming the Information Explosion

Within information lies insight – defined as the contextualized sum of facts, events and relationships obtained through algorithmic analysis of text and data from within and outside the enterprise – used to infer relevant and potentially non-obvious patterns and implications. We are creating a suite of innovative insight-generation services, integrated into platforms for harnessing the entire range of unstructured information types, at the scale of global enterprises and information providers. Research problems include coping with the vast scale of information sources, navigability between related information, and determining the importance of information, i.e. its timeliness and contextual relevance.