By Siobhán O’Connor:
As a fledgling researcher, I heard the term ‘Big Data’ several years ago when it appeared on the cover of New Scientist. It was billed as an exciting new field that was evolving at the peripheries of lots of disciplines and one that could potentially revolutionise them all. Having scanned the article briefly I didn’t make much of it at the time and resigned it to the realm of techies, one which would have little impact on me and the way I lived my life. However, as the years ago by and the proliferation of digital data seeps into every facet of life; from monitoring what I eat and the exercise I do via mobile apps, to sharing my personal data on family, friends and life events on social media, I realise I may have missed the central point of the article. The technology to continuously monitor human life (both biological and behavioural) and the environment that surrounds us is here.
When the opportunity arose to learn more about Big Data, I jumped at the chance, as I could see its potential for post-doctoral research. The Association of Commonwealth Universities (ACU) was holding a summer school on the topic of ‘Big Data and the Digital Divide’ for postgraduate researchers and I was lucky enough to secure a scholarship to attend. The weeklong initiative was held in Canada and brought together over 20 international researchers from several commonwealth countries.
On our first morning we got a tour around Toronto from our student ambassadors, before travelling to the University of Waterloo where we listened to interesting perspectives on Big Data from two industry speakers. Kevin Keane, co-founder of Brainsights, spoke about a range of wearable technologies like FitBits and Hexoskin that can measure a variety of physiological signs which are enabling us to quantify and understand different aspects of human behaviour. The following day, Mark Damm the CEO of Fuse Forward outlined how cloud-based analytics platforms can be used to process large, diverse datasets that can help us better predict a range of outcomes and impacts. That afternoon, Professor John Hirdes from the School of Public Health discussed how Big Data could impact the healthcare system if we used standardized clinical assessments to collect digital data and create more robust predictive clinical decision support systems for doctors, nurses and other health professionals.
On Tuesday morning, Dr Christian Boudreau spoke to us about how statistical techniques such as survival analysis are being used to query large amounts of bone marrow data. This is helping researchers to understand how long patients will survive if given a particular treatment or drug. A trip to nearby Wilfrid Laurier University to meet Dr Colin Robertson and his team resulted in us appreciating how geographical information systems can contribute to Big Data and help us understand the spread of diseases such as Japanese Encephalitis in Nepal. We were then treated to a real Canadian experience by going camping and canoeing at the nearby Laurel Creek nature reserve. Of course important data can also be gleamed from the natural environment, which heavily impacts our health. We were given a demonstration of a miniature drone or UAV (unmanned aerial vehicle) that remotely collects environmental data such as wind speed and direction, air temperature, agricultural practices and other geospatial data. After a campfire, sing-along and too many roasted marshmallows, we spent Wednesday morning canoeing around the reservoir with Dr Robert McLeman to learn how citizen science initiatives can help us monitor environmental changes in the weather as well as animal and plant species through mobile technology (http://www.ontarionature.org/protect/species/app.php).
For the last leg of our journey, we travelled to Western University in London, Ontario where Dr Mark Daley and Dr Dan Lizotte spoke about the difficulties of statistically interpreting big datasets. They highlighted the importance of robust algorithms for drawing correlations and causation, as mistakes can lead to glaringly strange results as shown by Google’s initial Flu Trends data which has since been corrected. They were joined by Dennis Buttera from IBM, who reiterated this point from the industry perspective. He outlined the challenges IBM face supporting different businesses from multiple industries with analytics capabilities. This in part is due to a lack of computer and data scientists and he highlighted the urgent need for many more skilled graduates in these areas. He recommended a book called ‘Thinking About Data’ to educate everyone from children to adults on Big Data, which is now at the top of my reading list (sorry pile of JAMIA papers sitting on my desk!). In relation to healthcare, he noted that precision medicine is taking off as people are using large DNA and genomic datasets to personalize drugs and treatments for individual patients and the potential benefits of this for society are still in its infancy.
As the week drew to a close and we said our goodbyes after a trip to Niagara Falls. As it turns I couldn’t have been more wrong about Big Data – it is here and it is going to affect us all as the wealth of digital information at our fingertips will only continue to grow and grow. For health researchers that means more multidisciplinary collaboration with industry and other research disciplines, both from the technical and social sciences, to broaden and enhance our understanding of how Big Data can improve health and wellbeing.