Just half a decade ago, it wasn’t economically viable to effectively sort and pull together all relevant data sets to address key business questions. But developments in big data technology mean businesses can now effectively make use of a growing range of data at a lower price point and in a more user-friendly form than before.
In particular, the process of information discovery, in which raw data is translated into business insight, has the potential to revolutionize business processes and innovation.
Many forward-thinking companies are already using information discovery and predictive analytics tools for relatively basic tasks such as forecasting fluctuations (in sales), but also for more advanced purposes such as predicting when machinery may fail or determining the impact of a change in working practice. In all cases, the insight gained is being used to improve efficiency, time to value and productivity.
Belgian media group De Persgroep applied predictive analytics to its customer data to gain better visibility of its newspaper customer base. By making use of Oracle’s Big Data Appliance to analyze customer behaviour, De Persgroep can now predict with 92 percent accuracy when a customer is going to end a print subscription.
Such is the level of information discovery provided by big data technologies, that instead of addressing a single question with a narrow scope or timeframe, analytics tools can now effectively empower an analyst to have a ‘conversation’ with the data to produce deeper and more sophisticated insights.
The process of information discovery, in which raw data is translated into business insight, has the potential to revolutionize business processes and innovation.
Information discovery is a key concept leading to these ‘data conversations’. In the process of pulling together a large amount of data to find correlations to answer a question such as “is a reader about to end their newspaper subscription?”, big data technologies can prompt and then answer deeper and more fundamental questions. Essentially, it’s about working with data in an agile way.
CERN, the European Organization for Nuclear Research, is making use of Oracle’s Big Data Discovery for analysing the performance of its particle accelerator, the Large Hadron Collider (LHC). The LHC is the largest and one of the costliest scientific instruments ever built and any insight into how to increase performance and availability is extremely valuable when one considers that the LHC generates 1TB of data per second.
Analytics tools can now effectively empower an analyst to have a ‘conversation’ with the data to produce deeper and more sophisticated insights.
Another example demonstrating predictive analytics’ power to detect patterns in seemingly random data can be seen in the financial capital markets, where it is being used to address the issue of rogue trading. The markers of rogue trading activity are numerous and disparate, making it difficult to spot suspicious activity in a timely manner. One Oracle client identified 35 markers, such as traders being less likely to leave their trading book unattended if they have done something untoward, or not handing over their book when they take time off. Changes in holiday patterns and signs of increased stress were also identified.
In this case, big data analytics was able to cut across data siloes to pull together information from different sources to predict when rogue trading might be about to take place, thus enabling organizations to take steps to prevent it.
Although predictive analytics isn’t necessarily something many businesses have in place, the technology is already available that will form the foundation for the next evolution of analytics: prescriptive analytics.
Essentially, prescriptive analytics combines predictive capabilities with automation in order to drive changes and actions within organizations. It enables organizations to streamline processes or drive improvements in operational efficiency without the need for human intervention.
When the insights of predictive analytics are applied through a rules engine, changes to processes can be automated. So if an event (either positive or negative) can be predicted, an action can be triggered to avoid or bring it about.
The technology is already available that will form the foundation for the next evolution of analytics: prescriptive analytics.
This approach can already be seen, once again, in the financial services, in the form of algorithmic trading that responds to changes in market rates to make or avoid certain trades.
But the applications of prescriptive analytics are broad and far-reaching: data generated by work practices could be analyzed to provide guidance to workers about how they could perform more effectively. Oil rig operators, meanwhile, could use information generated by sensors on the drilling platform to improve health and safety practices, maximize extraction from depleted wells or improve the effectiveness of maintenance.
Although very much in its infancy, the combination of predictive and prescriptive analytics is a concept that will become much more familiar to businesses over the next few years. And critical to this will be ensuring that big data technology, particularly analytics tools, can be easily upgraded as new capabilities come on stream.
As increasing numbers of businesses realize the value of predictive analytics and start thinking about the step into prescriptive, now is the time to be on the front-foot with the next evolution in data analytics. With the right technology and infrastructure now both readily available and affordable, the prospect of truly data-smart businesses has never been closer.
This Big Data article is brought to you by Oracle and Intel®.
Intel® and the Intel logo are trademarks of Intel Corporation in the U.S. and/or other countries.