Advanced analytics for a data enabled economy

4 min read

In the digital era, successful economies and businesses will be creative, innovative and economically diverse, driven by the generation and use of ‘Big Data’ created by computers, sensors and other digital devices, networked systems and improved analytics.

Concepts such as Industry 4.0 are being deployed and manufacturing companies are applying advanced digital approaches to improve their productivity, flexibility and quality.

The demand for greater interconnectivity is intensifying and, even amongst SMEs, there are moves to develop strategies to embrace more advanced concepts. Manufacturers are looking to embedded connectivity, Big Data, advanced analytics and data-driven services to better manage their information and, helped by the technologies associated with the IoT, are creating new business models.

The amount of data that organisations are looking to handle is growing exponentially. More data are being processed in real time and the variety of data has diversified. Crucially, there has been dramatic growth in unstructured data – but that is of limited value.

According to Jim Norton, chair of the Royal Academy of Engineering’s Digital Systems Engineering Community of Practice: “The idea of ‘Big Data’ is meaningless. I prefer to talk about data analytics – without being able to analyse the data, it becomes essentially useless.

“Data analytics comprises three elements: the ability to access large amounts of data – in many cases, collecting it in real time; having the processing power and the advanced algorithms to be able to do something useful with that data; and using that data to deliver better business decisions.

“We need to talk about smart data … there’s a lot of ‘rubbish’ data out there, which means it can all go horribly wrong if you’re not careful.”

Mandy Chessell, an IBM Distinguished Engineer who works in the company’s Analytics Group CTO Office agrees. “There are big questions around the concept of Big Data. Too many organisations are looking for a ‘silver bullet’, but there isn’t one! What you need to be is very systematic about collecting data and to be sure about what you are trying to achieve. You need to know what you’re looking for and your destination.”

How data are gathered is crucial but, more importantly, so is the metadata that is generated.

“Metadata is the key,” Chessell argues. ”As you pull data from different systems or sensors, you need to capture as much of the ‘tribal’ knowledge that you can – the metadata. What is the sensor really measuring?”

Metadata is the base for harnessing data generated from disparate data sources and information repositories before it becomes unmanageable. Without metadata, it is suggested, firms will not be able to extract the deep insights that Big Data can yield.

Metadata ensures a more accurate picture of data and greater data consistency for Big Data analytics and business applications.

Interestingly, metadata rarely comes up as a critical priority, according to a recent piece of research from Gartner. Rather, the focus tends to be on unstructured data content – which could represent as much as 80% of a firm’s total information assets.

“While Big Data isn’t new,” explains Chessell, “what has changed is the varieties of places that are now creating data. Companies and organisations are no longer in control of all their data, most are using data from external multiple sources.”

Today, data analytics is becoming complex and aggregating data from multiple projects and/or applications is an increasingly intricate challenge.

Single analytics platforms embedded within core operational systems are becoming more popular as they help to ensure that key data are integrated.

“It is crucial that data are not stored in silos, where their insights will be lost,” says Norton. “We need to find ways of sharing it and of making it actionable, but within a legal environment. If we can liberate that data, we will create value, but we currently lack the tools sets necessary to deliver this.”

Norton highlights the approach of the UK’s Digital Catapult, which works across a range of technology layers to deliver better data management and analytics, as well as the need for infrastructure regulators, professional institutions and standards bodies to work together to provide better codes, guides and specifications.

According to Norton, over the past 20 years, we have seen a growing move to what he describes as ‘a data enabled economy’. “The rise of the Internet has been key to that; now the focus is on turning data into information or smart data,” he says.

“To succeed, organisations need a better understanding of the strengths and weaknesses of data analytics. It is quite possible to take a set of data and find a pattern, especially if you want to find one. For analytics to bring real value, it has to be done vigorously and, unfortunately, we don’t have enough people with an instinctive understanding of data generation.”

“We need to improve educational and professional development and encourage data science skills,” Chessell suggests.

There are some success stories though. Norton points to companies taking a product and turning it into a long term service though analytics.

“Whether companies are doing data analytics well is moot, but if you take manufacturing, for example, companies like Rolls-Royce no longer make jet engines, rather they sell power by the hour. What they offer, through the collection of data in real time, is pre-emptive maintenance and a more efficient long-term management of jet engines.”

Another example is Network Rail which, according to Norton, is ‘doing some interesting work in predictive maintenance and knowing where its assets are. “There are good pockets of best practice across the UK; we just need to spread that knowledge.”

Mandy Chessell

Norton believes that, by linking different data sets, we will see the value of that data go up enormously.

“But that will depend on the introduction of standards for data and metadata and that has not been properly developed yet. It also raises the important question of how you value data and how it is controlled and tracked up and down the supply chain.”

The IoT and the deployment of sensors is opening up many new possibilities, suggests Chessell. “Traditional business systems tended to be built and designed around the progress of that organisation. These types of organisations use Big Data to provide a coherent view of their operations, enabling them to work with new technologies. It provides them with a mechanism to connect to their existing business and to rationalise existing operations. Banking, retail and insurance are good examples.

“The development of IoT sensor technology is opening a host of new possibilities. Companies are now exploiting their IP and we will see massive disruption across many traditional industries as new players, with new insight and value, look to use resources more effectively.

”We are seeing a trend which is looking to push processing power closer to where data are created. Data needs to be managed and protected throughout its lifetime. If data are to be valued, it needs to be secure, trusted and you don’t want to lose it.”

“We need to understand how we gather information,” Norton concludes,” and calibration is vital. Is the data captured accurate? Is it real? If it’s not the whole premise of ‘Big Data’ is undermined.”