Sunday, 20 September 2015

Data and Unlikely Academic Collaborators

Last week at a social event I met two interesting academics, working in apparently diverse fields –chemistry and social science. It became clear in conversation that academic distinctions mean much less than they used to, and rather than becoming more specialised, many research disciplines are expanding to such an extent that boundaries are blurred and perhaps even irrelevant. Both these academics, for example, use computational methods not just to aid their research, but at its core.

Working in Cambridge exposes one to all sorts of academic research and thought – it is ceaselessly fascinating to meet real experts in their fields from all over the world and to hear first-hand about the early Christian temples of the Middle East, developments in cancer prevention or the history of concentration camps. In today’s cross-disciplinary environment one can risk overload when considering the multitudinous connections between research disciplines – particularly when working, as I do, in a field which facilitates collaboration. For decades, ‘hard sciences’ such as particle physics have generated massive amounts of data, and developed advanced statistical analysis tools to cope. It didn’t take long for financial institutions to grasp the power of these mathematical techniques and computational approaches (and to hoover-up PhD graduates), but it has taken longer for other academic disciplines to catch on. In the last ten years or so, there has been a significant change: highly data-centric scientific disciplines boomed – computational biology being an excellent example, and specialist hubs are no longer solely at traditional centres of excellence like Cambridge but are truly global. Furthermore, and perhaps less expectedly, big data has been embraced by social sciences, humanities and even the arts.

The cynic might see some of this activity as a reflection of the poor state of research (and particularly arts) funding causing academics to jump on the latest ‘data’ bandwagon, regardless of its relevance to their work. This is undoubtedly true in some cases, but there is real academic rigour in much of this work; surely it is not so surprising that statistical analysis on a massive scale should return to its home territory of monitoring and analysing human society – just as in the Doomsday Book. What is more, new swathes of cross-disciplinary collaboration are opened up: the cross referencing of data sets and scientific models from disparate sources can lead to radical and sometimes counter-intuitive findings. This trend is taking hold across academic study encouraged by forward-thinking organisations such as GEO in earth-observation, mapping land and sea measurements against atmospheric and satellite data, and Cambridge’s own CRASSH in the social sciences, whose cross-disciplinary seminar series can be amazing to attend.

The scope and form of the research groupings that will emerge from this collapsing of boundaries is difficult to predict. What is already clear, however, is that these new fields have specific requirements which go beyond traditional data handling tools. Technologies which allow metadata to be cross-referenced across academic boundaries, methods of uniquely identifying research findings, and secure data-sharing technologies will all become key to the next generation of scientist-cum-data-beachcomber. What isn’t clear yet, is who will establish the universal standards in this space.

Sunday, 19 July 2015


It is clear from any newspaper or news website that wars don’t just happen anywhere. It seems that the scenes of unimaginable suffering often occur in countries where, if the news spotlight wasn’t attracted by the violence, some pretty everyday suffering would be going on anyway. To be highly simplistic, wars generally happen in places which are poor. You might find a cynic somewhere that will tell you that this proves that poor people are intrinsically more violent or that violent people tend to be poorer; but any rational soul will conclude that scarce resources can often lead to unrest and even war. This isn’t just about competition for the slender pickings of an unforgiving environment – lack of prosperity or unevenly distributed wealth leaves populations with little ownership and hence no personal investment in maintaining a peaceful society. 

On Tuesday, I attended International Alert’s event ‘Peace through Prosperity’ at the Royal Society in London. This provided an interestingly balanced perspective chaired by Alert’s Dan Smith with a panel comprising experts from industry and journalism as well as the INGO sector. Of course, given the topic in hand, there was little in the way of combative debate, but a range of experience was drawn upon and provided diverse perspectives on the subject. Particularly pertinent to me was the discussion of infrastructure development – and its potential impact on regions’ abilities to prosper. Naturally, experience is largely focussed on water provision and road building, but equally relevant in the 21st century, if we are to focus beyond subsistence towards prosperity, are electrical power and internet connectivity. The author of the report which prompted the event, Phil Vernon, rightly highlighted the case of Rwanda, in which a recently war-torn (and still not entirely peaceful) country, has focussed on developing an IT industry as a means of modernisation and nation-building. Internet connectivity, now such an intrinsic part of western economies, brings significant benefits to otherwise isolated regions. It can allow local businesses to become global in their outlook, students to gain a world-class education or researchers to access data from cutting-edge experiments only milliseconds after it is uploaded. The long-term economic and peace-building effects of such advances will be extremely interesting to monitor. Africa certainly has the opportunity to develop such infrastructures –there are now sea cables on both east and west coasts of the continent, connecting to major internet hubs in Europe. National infrastructure may be harder to fund and deploy, particularly for land-locked states. However, in the research and education space, the Ubuntunet Alliance is entering the second phase of the AfricaConnect project, where predominantly eastern and southern African NRENs (National Research and Education Networks) –including Rwanda of course – work together to facilitate international network access. In a continent which has seen a disproportionate lack of prosperity, it must be hoped that such developments bring not only information, but peace.

Wednesday, 24 June 2015

Academic partnerships building development

Crises in countries on the fringes of Europe dominate world news: Ukraine, Syria, Libya and Egypt have replaced Kosovo, Georgia and Lebanon in the headlines in past years. What these countries have in common is not only border and ethnic conflict, but a limited ability to develop advanced economies due to (amongst other factors) a crippled and underfunded education and research sector. This leads not only to an isolation from global scientific communities, but to a generation of students without the skills needed to build the high-value SMEs and start-ups essential for a vibrant and globally-engaged economy.

It is well known that the developing world suffers what has long been termed the ‘digital divide’. This term has typically referred to the historical lack of availability of internet bandwidth in geographically isolated countries in the grip of incumbent telecommunication monopolies. Although this situation persists in many countries (even in some EU member states), the price of Internet connectivity has plummeted in a number of countries previously badly impacted –Ukraine and Turkey are two examples where wholesale internet connectivity is available at less than €1/Mbps/Month –comparable with pricing in Western Europe.

So, it is certainly encouraging to see widespread access to the Internet –but unfortunately this does not equate to good access to scientific data and facilities. Open Data is a hot topic in certain sectors of the scientific community, but its impact will not reach beyond leading scientific nations if high-quality infrastructure is not available for research and education. This infrastructure comprises not only internet networks (at a higher quality and capacity than commoditised services) but also computing and data storage facilities.

The European Commission has recently signed an €13.7M contract to deliver a high speed academic network for Eastern European and the Caucasus –based upon feasibility work conducted in my team at G√ČANT. This will go a long way towards building essential capacity –both technical and human –but the focus needs to be wider. CERN is building partnerships in developing regions –notably for this discussion, the Caucasus and North Africa. This promises to begin to spread the benefits of global scientific collaboration to developing countries –in terms of access to the world’s most advanced (and expensive to produce) scientific data, but also in beginning to build local data infrastructures to exploit the data. This is a great step forward, but if genomics, climate and agricultural data can also be accessed through the same pathways, and locally-relevant analysis can be computed, then the societal impact will be more immediately tangible. Given the recent dramatic flooding in Tbilisi, the relevance of data-centric disciplines such as climate-modelling cannot be in doubt.