In May, The Alan Turing Institute hosted a week-long Data Study Group event organised by Turing Faculty Fellow Sebastian Vollmer: an opportunity for academics to work on data science challenges posed by industry. As a Turing Doctoral Student, this was my second time taking part in a data study group, and on this occasion, I was a facilitator for the Defence Science and Technology Laboratory (Dstl) – Urban Analytics group.
Initially, I was a little nervous about guiding a team of mainly strangers (the data study groups involves academics from all around the world) to deliver some useful methods for Dstl to take forward in less than a week! Luckily, unlike a hackathon, we were not obliged to fully complete a solution to tackle the problem. In fact, the data study group’s intention is to provide companies with several methods they can take forward either in-house or as a future collaboration with the Institute.
I think this is great as although hackathons are fantastic at generating innovative solutions quickly, due to the time constraints often ideas haven’t been thoroughly tested and then fail at the implementation stage. Instead, providing industry clients with cutting-edge academic knowledge on how to apply different data science approaches to tackle their problem is far more valuable, as there is real knowledge these companies can take back and apply to other challenges the companies might face in the future.
The project that Dstl proposed was a very interesting challenge to tackle. We analysed large amounts of spatial data, such as highways networks, points of interest and locations of crime activities, in order to develop models which can predict likely locations of facilities (e.g. nuclear power plants) or events (e.g. civil unrest) at a global level.
The challenge was particularly interesting from a research point of view as there could be many different approaches to solving this problem. Therefore, the project drew a fantastic selection of candidates to our team from different fields – varying from spatial analysis and epidemiology to mathematics and physics.
In the first few hours, we produced a plethora of ideas on how to tackle the challenge Dstl had posed to us. We decided to split into subgroups and explore our different approaches, with daily catch-ups to update the group on our progress. This didn’t mean that the subgroups purely stayed working on their own approach: as soon as the other groups saw how the other approaches could enhance their own approach, they started to include these insights in their own work. I saw some great interdisciplinary collaboration.
Dstl were extremely impressed with how quickly we managed to come together with so many approaches to tackle their challenge, as well as with the dedication of the team members to find useful solutions.
Matt, a Dstl representative, commenting on the work developed, said:
“The Data Study Group week was one of the best experiences I’ve had working with external suppliers of any kind in my 8 years in the Ministry of Defence. A large team of very capable academics worked extremely hard on my problem and demonstrated multiple approaches to solve it. To have so many interesting approaches thoroughly explored in such a short time period is very impressive and incredibly useful to our organisation. The team was very well led by the Turing facilitator, Chanuki Seresinhe, the few IT issues that arose were dealt with immediately and on occasion world leading experts in particular fields were called in from elsewhere in the Institute. I have already recommended to colleagues that more of our problems should be pitched to the Turing Data Study Group events.”
Overall, this was a fantastic experience. If you are an academic who wants to get some hands-on experience working on industry challenges in a dynamic and inspiring atmosphere, then I would highly recommend joining a future Data Study Group. I also think the Data Study Group is equally valuable to industry partners, as it is rare to be able to work with a group of top academics around the country all at once for some valuable insights into cutting-edge approaches and also just some fresh inspiration.
The Alan Turing Institute organises several data study groups per year. If your organisation would like to get involved please email email@example.com.
This blog was written by Chanuki Seresinhe, Turing Enrichment Year Doctoral Student and facilitator for the urban analytics group which worked on challenges posed by the Defence Science and Technology Laboratory (Dstl).