By Niall O'Doherty, Head of Business Consulting at Think Big Analytics, a Teradata company.
For decades, many industries have integrated data and advanced analytics into their business and digital transformation strategies, enabling them to better inform the decision-making processes and drive competitive advantage.
The manufacturing industry is no exception, however there seems to be a big gap between the leaders and laggards in capitalizing on opportunities presented by advanced analytics and, in particular, sensor data and the Internet of Things (IoT).
Large, forward-thinking manufacturers are using integrated data and advanced analytics at scale to build a more holistic, detailed view of their organizations and to turn the large volumes of data into tangible returns. Investments in IoT platforms are becoming increasingly sophisticated, and the forerunners are recognizing the potential; the ability to feed data into the advanced analytics lifecycle, optimizing efficiency and productivity levels across entire value chains, while reducing overall costs and inefficiencies.
So, what's next? How will advanced analytics and IoT progress to help manufacturers see further business benefits in 2018?
IT/OT: integration to drive business outcomes
Integrating sensor and non-sensor data, manufacturers will deliver transformational business value. While analyzing sensor data on its own provides some limited value, the future for the manufacturing industry will rely on the integration of the sensor data with other types of data (customer, product configuration, financial, macroeconomic, etc.) to achieve more sophisticated results from analytic techniques.
Increasing data integration will result in the merging of IoT platforms with business platforms over time. In technical terms the consolidation of the Information Technology (IT) world (ERP, CRM) and the Operational Technology (OT) world (shop floor MES systems, telematics). At a basic level, the process involves aligning the two disparate worlds: the OT talking about protocol level data integration, and the IT side talking about big data, traditional ERP, and data integration.
To get the two talking the same language regarding combining data sets will be a challenge. However, it's essential to making progress in the manufacturing industry: informing better business decisions, optimizing business processes, reducing operational costs, driving competitive advantage and delivering growth. Our goal is to deliver this integrated IT and OT platform capability for the enterprise, especially where a platform does not refer to hardware but rather a capability that is inclusive of people, process, and technology to achieve agility.
Multi-genre analytics enabled
To succeed in an increasingly digitalized industry, we see manufacturers attempting to deliver multi-genre analytics - i.e., performing multiple analytics capabilities (using everything from text analytics and statistics to Artificial Intelligence and Deep Learning) against the same data. Multi-genre analytics allows organizations to extract value from new and old data, enhancing understanding of a huge range of business-critical events surrounding customers, products, and interactions.
We've helped the US Army increase the readiness of its helicopter fleet by between 5 and 8%, for example. We've helped Union Pacific Railroad cut wagon bearing-related derailments by 75%. We help one of the leading manufacturers of paper mills avoid the mother-of-all paper-jams in its mile-long plants. We help keep the lights on by predicting failures in the electricity distribution network before they occur. And we help a European train operator figure out when trains will fail up to 36 hours before they actually do.
The interesting thing about these different projects is that despite their apparent diversity, from 50,000 feet they mostly look remarkably similar. These projects depend on the availability of relevant sensor data, including temperature, pressure and vibration measurements to name a few. However, we typically need to pre-process the raw sensor data to identify changes in state – "events" - which requires time-series analytics. We typically need to label the sensor data in order to train a supervised predictive model, which often requires that we use text analytics to extract fault and resolution details from engineering reports - and may additionally require us to join the sensor data with operations data.
When we have labelled event data, understanding the sequence of events that leads to a particular condition (the "path to failure") is normally important – which requires path analytics. And because understanding associations and relationships – between different events and different components – is often vital, we typically also need to apply graph and affinity analytics to understand which variables are likely to be predictive of the target that we are trying to model. Multi-genre analytics!
The predictive models that underpin many analytic IoT projects are often relatively simple. But the process of creating a useful Analytic Data-Set from raw sensor data is often anything but simple. Enabling multi-genre analytics at scale will provide clear insight and allow the manufacturer's analytical maturity to increase, providing companies with the opportunity to answer more complex questions, based on a broader, more robust range of evidence.
Analytics Ops for scale
The third critical requirement that needs to be addressed is delivering the data integration and analytics at scale in an operational environment that will enable continuous improvement. Organizations need to build a lasting analytic capability. As organizations move from research and development projects to running analytics at an operational scale every minute, every day, without fail there needs to be a bridge from the Proof of Concept phase to Sustainable Data Products. Analytics Operations (Ops) is that bridge. It is all about production and sustainable development, the ability to Think Big, Start Smart, Scale Fast.
Analytics Ops is focused on delivering business value, which comes only when a data product is in a production state. Bringing products to production is a process, and maintaining them when they are in production is also a process, requiring process engineering as well as data science and data engineering involvement to effectively integrate into existing business systems.
Analytic Ops is a modern Best Practice framework aimed to operate production-grade analytics across cross-functional teams leveraging proven best practices from the world of software engineering. It breaks down silo thinking and creates enterprise-ready data science following a DevOps model for Data Science. Operate, maintain and improve your advanced analytics so you can transform your manufacturing capabilities. Whether it is manufacturing toys or trains, we have helped companies implement Analytics Ops so that they can stay ahead of the competition.
To 2018 and beyond
The 'factories of the future' are already emerging. Whether it's car manufacturers investing in plants that can create multiple vehicles off the same production line, or semiconductors labs reducing millions of variables within complex structures to identify critical variables affecting yields, IoT and advanced analytics show no signs of slowing down anytime soon.
As with 2017, 2018 is set to be a year of evolution in the manufacturing space, where companies will continue to use these technologies to analyze data in real-time and to use this data to scale to see significant business benefits faster than ever before.
As the "factories of the future" emerge Teradata is working to deliver the technologies, expertise and capabilities that will allow large manufacturing corporations to deliver on their promises towards an even more exciting future for the enterprise. Our efforts today focus around creating agile and transparent data architectures and systems will enable us to deliver integrated data, multi-genre analytics on a scalable platform through Analytics Ops.
At Teradata, we work on helping manufacturing companies to sense when something's wrong and report it to the humans in charge of fixing it. In the Sentient Enterprise, an entire company operates like a single organism, where the left hand knows what the right hand is doing, and where human beings can get signals and suggestions that inform and guide their critical business decision. It is this systems of systems view, and not just the "things" view in the Internet of Things, that will deliver transformational change.
 "Sentient Enterprise: http://www.teradata.com/Press-Releases/2017/The-Sentient-Enterprise-New-Business-Book