Industrial Data Summit 2018: The big data roundtable talk

Posted on 23 Apr 2018 by Jonny Williamson

The Industrial Data Summit brought together 100 decision-makers in the Mary Ward House, London, on 18 April. In interactive roundtable sessions delegates discussed how businesses could leverage new digital technologies.

The Industrial Data Summit brought together 100 delegates discussing how digital technologies could leverage the business – image courtesy of The Manufacturer.

It was probably the UK’s most informative data summit of its kind, The Manufacturer’s Industrial Data Summit 2018.

Around 100 decision-makers across all manufacturing industries gathered on the first sunny day of the year to discuss how data can be turned into an enabler of business growth.

Nine roundtable discussion sessions formed the heart of the Industrial Data Summit; and the idea of this conversation format was as simple as effective.

The delegates could pick out five different subjects throughout the day to discuss a variety of topics around industrial data, such as IoT, IT/OT convergence, machine learning or AI.

I joined the three discussion sessions on AI, Data & Customers and Identifying business data problems in general, which was a good chance to gather more information about big data and the possibilities to turn information into money.

AI & the Industry

Ever more companies plan to adopt or have already deployed AI in their factories; this has become almost a matter of fact.

At my first roundtable, AI for Industrial, delegates took the opportunity to exchange their own specific experiences on using and monetising business data with the help of AI.

Delegates could choose a 30-minute discussion from across all three streams – image courtesy of the Manufacturer.

It was interesting to hear that even large companies are still struggling with very basic processes, like transforming information written on paper into data which can be processed digitally.

A delegate from a metal component manufacturing SME said: “It was good to find out that we are not the only company having just started the journey.

“And it was reassuring that even some of the big companies are struggling as well and they are also on their journey without being quite sure where to go.”

On the other hand, the roundtable discussion revealed that there are already many manufacturers out there actively deploying AI on the shop floor.

A director from a major car manufacturer said his company was already deploying digital twins to make the production processes more efficient. And, also glasses are set to play an increasing role on the shop floor, especially to teach and train staff.

A summarising sentence at the end of my first 30-minutes session could be: It is always good to work out the potential benefits and the outcome of a project first, before you start deploying AI.

Demystifying the process

Defining and identifying a data problem type and turning this into a business advantage by monetising data through services, might be one of the biggest challenges a company’s digital transformation has to face.

The second roundtable I joined, From Business Problem to Data Solution, revolved around the question: How to extract value through the definition and application of data solutions?

Discussion host Nick Frank kicked off the conversation with a punchy sentence: “Forget the technology in the first place and begin with a business plan!”

And the second session host, Eric Topham, underlined that most important in the process of identifying a data type problem is to start with a hypothesis.

Topham said that the best way to define and identify a data problem is to answer the four V-questions.

  • Volume (How many exabytes, gigabytes?)
  • Variety (Is the data structured or unstructured?)
  • Velocity (What do we know about the frequency of the incoming data?)
  • Veracity (Is the data trustworthy at all?)

Considering the four Vs, Topham said, can help a company formulate a concrete hypothesis, which can better describe a company’s internal ‘data chaos’.

Topham concluded, that the first part of getting a company’s mind ready for the digital transformation is to set up an (at least) testable hypothesis based on the four V-questions.

Only then, the big data journey has a realistic chance to kick off.

Collaborating with customers

Businesses produce data, and collecting information has become a daily business routine, it seems. But the question is: What does a business have to consider when it comes to sharing this data with customers?

Marco Del Seta, host of my third round-table, Collaborating with Customers, said right at the outset: “We live in a sharing economy. Everyone has to share, it is almost like an obligation.

“But how much data do we really want to share with our customers. This is a crucial question, also from a risk perspective.”

Another key issue, when it comes to sharing data with customers, is clearly the data quality.

Philip Woodall, researcher from Cambridge University, explained: “It is all about metadata; how can a company maintain the data quality and keep control over the data transformation process?

One key issue, when it comes to sharing data with customers, is clearly the data quality – image courtesy of The Manufacturer.

“A business has always to ask itself, which data should be transferred, and which shouldn’t and in which form.”

Furthermore, the delegates discussed problems occurring around interacting with customers through a vivid data exchange.

A director from pharma firm said: “A company might not feel comfortable with the idea of sharing data with other clients – just because of security and data protection reasons.”

He explained that sometimes it is difficult to get data from a customer’s machine although it would help him to improve the production process.

“If you are working in the pharma industry, there are companies which will not allow you to plug their machines into the internet to get data.

“Although, it is our job to get our clients’ machines to make some clever things, the clients themselves are not very comfortable with us opening their machines up to data exchange processes.

“They invest for instance a lot of money in developing new proteins, and we make machines that help to improve their performance.

“But, once you link into their system, they think somebody else might hack into it and steal information; and this is a security issue. We have realised that we are all very confused about the security aspects.”

Our upcoming summits will be: