Прогнозная аналитика для Больших Данных

Yesterday’s Analytics

Until recently, financial analysis-based decision making relied on a historical view of aggregate data – understanding yesterday’s events to influence tomorrow’s performance. Highly skilled data scientists took summary-level information – analytic engines would be quickly overrun by larger or more detailed volumes of data – and analyzed promotion effectiveness, capital ROI, or quarter-end results. While useful, this approach posed several problems.  Analysis was by definition backward looking. Summary-level analysis could easily mask anomalies that, if identified, would change the course of a decision. And because analytic tools were so difficult to use, only a few specialists could use them, which created bottlenecks and limited dissemination.

Today’s Analytics

Today, enterprise data is growing exponentially and raising questions on how to manage big data for enterprise value. SAP HANA has broken open analytics with its powerful in-memory computing, making it possible for businesses to analyze large volumes of data in real time at the most granular levels. Easy-to-use tools put sophisticated predictive analytics, once the domain of senior business analysts, into the hands of employees across the organization. The result: More people get deeper insight more often, increasing their ability to affect business performance.

Predictive Analytics in Practice

Let’s look at how the power of predictive analytics has affected a wide range of industries:

• Power grid monitoring. A European utility operator radically improved transformer monitoring by loading readings from more than 22,000 sensors in 400 transformer stations into an SAP HANA database. Now, with new data streaming at five-minute intervals, statistical algorithms automatically detect and address data quality problems, avoiding overloads. SAP Business Objects dashboards and reports let users visualize and analyze the cleansed data, minimizing manual effort while increasing transparency.
  
  
• Bank fraud detection. Banks combine predictive analytics with SAP HANA to identify fraud and avoid financial damages related to illegitimate activities. A combination of real-time rules, methodologies, and predictive analytics identify accounts and transactions with high potential for fraud, triggering increased surveillance and investigation.
  
  
• Credit card analysis. Another valuable use for predictive analytics in banking is credit card analysis. Banks can analyze millions of transactions per month, revealing geographic, shopper, and customer profile trends. The resulting information can be used to negotiate fees and charges and to inform future financial, marketing, and customer service operations.
  
  
• Hardware retail pricing. A large hardware retailer planned to attack its primary competitor by offering the lowest prices on every item. After running sophisticated predictive analysis and simulation, however, the company determined that reducing prices on a select subset of items was a better long-term strategic decision, radically improving overall profitability.

Source:  sap.com