March 23, 2017, by Stuart Moran
Research software engineering support in Geography
The Digital Research team are working with the School of Geography to trial a new Research Software Engineering (RSE) service. Our RSE has been across a range of different research projects including:
Doreen Boyd, Associate Professor and Reader in Earth Observation, and her team are using a series of rainforest based LiDaR data to determine the distribution of tree species. Their challenge has been a lack of sufficient computational power and in determining the best approach to analysing the full wave form. Our RSE has been helping to create scripts to automate the transfer of data between different software components and also run the data through the University’s HPC, in addition to advising on different analytical techniques. The windows environment created a lot of headaches around the potential HPC use, and we are currently looking at cloud computing options.
Climate and Ecological Patterns
Adam Algar, Assistant Professor in Biogeography and his team are looking to model to predict how climate can affect ecological and evolutionary patterns, particularly in lizards and frogs. In light of high barriers to entry for the HPC, the team and their research are limited by the compute power of their local machines. The RSE is helping to make accessing the HPC easier, and to also help them explore the power of scalability in the cloud which may create new opportunities for their work.
UK Bank Closures
Andrew Leyshon, Professor of Economic Geography and his team have gathered a unique set of longitudinal data over the last 30 years on bank closures throughout the UK. Through the RSE, an opportunity was seen to explore the role of agent based modelling in relation to this data. The plan is to create a model of consumer and bank behaviours, and to use cloud computing to run thousands of simulations against the longitudinal data. This will help to create a predictive model for exploring future economic scenarios.
Next blog in series: Working toward the provision of cloud compute services
Stuart Moran, Digital Research Specialist for Social Sciences