Big Data is Multi-disciplinary – by Chris Nott

UK Academics collaborate at IBM Colloquium on Big Data challenges

Chris Nott, CTO Big Data & Analytics, UKI; TCG Chair

 

When engaging with academia for big data, it isn’t just the computer scientists that have a contribution to make. IBM is engaging the attention of social sciences, the arts and humanities as well as the experts in science and engineering. Developing use cases for big data beyond the traditional sciences is a priority as we apply analytical techniques in new ways across research and teaching. Business needs a cohort of graduates with big data and analytics skills alongside a wide range of academic disciplines in order to drive not only economic but also societal benefits.

IBM’s Technical Consultancy Group (TCG), the UK and Ireland affiliate to the Academy of Technology, hosted 40 leading academics in big data and analytics from UK universities for a colloquium held last Wednesday at the Royal Academy of Engineering in London.

The purpose was to jointly explore three broad themes in pursuit of these benefits. The themes spanned research, teaching and technology in order to begin new and develop existing conversations between academic experts and leading practitioners from IBM.

 

• Applications and Adoption.
  • Vicky Brock, CEO, Clear Returns, described the commercial perspective on building a big data and analytics culture. Her need is for desire and creativity in her people to exploit data so that retailers retain more revenue post sale and improve each customer’s experience to increase overall lifetime value. Essential to success is having a culture which questions rather than procrastinates and is able to accept a degree of failure.
  • I described how businesses generate value when they act on insight from big data. Until then they incur investment cost to gain access to and insight from data. By taking an integrated architectural approach, big data and analytics improves business processes. One of my case studies described the social benefits derived from analytics.
  • Lauren Walker, big data analytics leader, IBM UK and Ireland talked through examples from many industries where IBM’s clients have implemented big data and analytics. They included using analytics to create new business models, attract, grow and retain customers, and optimize operations. Doing analytics well can make a big impact not only for businesses but also for their customers.
• Social and Media Issues
  • IBM Distinguished Engineer Professor Mandy Chessell highlighted some of the issues surrounding ethics of analyzing big data, especially personal data. Whilst consumers may be given a choice to receive a discount or reward in return for providing that data, they could see themselves being penalized if they don’t. There are many perspectives on ethics—and ethical judgment is personal—and companies perceived to be operating unethically can damage their reputations. Mandy described an approach developed by the TCG to help organizations develop an ethical framework.
  • Professor David de Roure, director of Oxford e-Research Centre, described how new forms of data allows researchers to pose questions which couldn’t previously be addressed. When using social media, for example, data collection is not required to explore areas such as food and energy consumption. We can observe the scale of a social process in real time with Twitter which couldn’t be done before. The biggest challenges lie at the intersection of open data and personal data and a multi-disciplinary approach is required to tackle them. For example, social science ought to include the social media itself, not just provide a lens on to what’s happening. He encouraged big data research to be “in it,” immersed in what is being explored, rather than “on it,” simply observing from outside.
• Research Challenges
  • Colin Shearer, IBM global executive for advanced analytic solutions, with Professor Dave Robertson, head of the School of Informatics at the University of Edinburgh, acknowledged that the era of big data has triggered such demand for insight that analytics needs to be industrialized. Observations on the agricultural and industrial revolutions show us that with big data we need to move from having artisans using ancient techniques, such as code, to industrial strength tools and systems that automate the analytics process and make it available to the masses.
  • A key role of analytics tools is to remove the bottlenecks on the scarcity of highly skilled individuals—automation allows us to scale analytical projects. One goal would be to shift from needing to hand craft every predictive model to focus more effort on improving the processes which use them. Furthermore business use analytics benefits from pragmatism; the goal is to be more right more often in operational decisions to generate a return and avoid spending time perfecting insight. It is foreseen that social computation can complement traditional approaches to tackle new types of problems that emerge. One of the research challenges is to understand the infrastructure that is needed to support this social approach.

In the afternoon, each of the three themes was discussed in open forums to further develop the ideas presented. These will be the subject of subsequent blog posts.

Many of the attendees expressed an interest in the business analytics module IBM has developed for Universities. It is currently being piloted at Manchester and Warwick Universities and is primarily aimed at Business School students. It starts by teaching the foundations and builds on these with the latest advanced in predictive analytics, big data, real time analytics and cognitive computing.

IBM has opened up its cognitive capability to build an ecosystem around our Watson technology. It offers further opportunities for universities to collaborate on this new frontier.