Over the summer break, 11 of New Zealand’s aspiring computer science students took part in the first Summer of eResearch to be held here.
Based on the Summer of Code concept, the aim of the Summer of eResearch is twofold: to help New Zealand science groups meet their computing needs, and to introduce computer science students to the scientific research being carried out. The result? Building mentor relationships between computer science students and our scientists, and to foster increased collaboration between the fields.
After a feverish few months, the students presented their work in late February.
Overall, the projects were aimed at a few, simple issues, many focussing around the ‘TMI’ (Too Much Information) problem. Given TMI, what does one do with all of it? How does one store it (in terms of format, etc). How does one input it, share it, and analyse it?
The range of challenges (and solutions) was impressive: from visualising drug discovery analyses to cataloguing model ecosystems, computation over geospatial datasets to genetic marker design workflows, as well as improved means of accessing BeSTGRID’s services.
Some other common themes emerged during the presentations and resulting discussions: the need for support from smart computing people, the enjoyment of the students and the synergies which spring up out of this sort of collaboration.
Need for support from smart computing people
Traditionally summer scholarship focus within one area of science, while many of these students worked with mentors in a science area along with advice and support from computer science and software engineering mentors. One of the primary issues around work of this sort is the need for continuance of the work being done. A great deal of work was done by the students and their mentors, but much remains to be done if the issues worked on are to be solved.
Ongoing mentorship for students who are interested in these challenges is vital if they are to remain interested and involved in the work.
Additionally, funding is of enormous importance. Software development and maintenance needs to be specifically funded, and funding is also required to turn proof of concept work into something which is viable.
Such funding, however, is not necessarily as high as people might fear: as an example, the cost of finding one stoat on Stewart Island would have covered the cost of 6 students for a project such as the Summer of eResearch. As these projects attest with a $5,000 scholarship funding each student for the summer, a little can go a long way
The enjoyment of students engaged in science
The enjoyment felt by the students as a result of their engagement in research science was palpable. They were able to see the real, and immensely valuable, contribution their skills could make to research in New Zealand.
Perhaps the best way to show the enjoyment felt by both mentors and students is to quote them directly.
Says Rachit Bhatia, a student who worked on Invader Genetics for New Zealand Conservation, a project aimed at improving bird conservation by centralising data collected by volunteers and researchers and hence also allowing ease of access:
“I had a great time working with SoeR (thanks Nick and Rachel). My project was very interesting and I loved the experience. I got the opportunity to take ownership of the project right from the beginning and I was granted enough freedom to take the project in the direction I was interested. As a student moving into the industry, this approach allowed me to gain highly needed skills and valuable analysis and development experience. Typically in the industry, it would take people a few years to move through as many phases of a project as I did within one summer. I believe that my time with SoeR has made me a better software professional at my new job.
“I would highly recommend the Summer of eResearch to any students looking to gain interesting, relevant and valuable experience.”
Rachel Fewster, the mentor for the Invader Genetics project, commented:
“The Summer of e-Research happened at just the right time for me. I was looking for ways to develop a database and GUI for invasive pest management. The database would be a one-stop shop for everything from fieldwork records to DNA analysis results. The GUI would link the data to maps, highlighting pest hot-spots, mapping changes over time, and locating where invaders have come from using DNA results.
At first I thought that I would just need a "simple" database, but I soon learnt that no database is simple! Through the Summer of e-Research, I was sponsor for two software developers (Sunil and Rachit), and gained a good understanding of what's involved in planning and developing a data management system. The planning phase was very careful and thorough, and it was a real delight seeing how the careful planning came together into the end product. We were aiming for a system that would be suitable for diverse projects, from long-term community conservation projects through to specialist scientific research. Aiming for this diversity, and having user-friendliness as an uppermost requirement, made the design quite challenging. I was surprised at how rewarding it was to think through all the possible requirements and create a catch-all data model.
Overall I'm thrilled that I had the opportunity to set up this system, with the confidence of expert technical input from the BeSTGRID team, and a truly professional approach from the two developers. All I had to provide were "needs" and coffee, and the rest was taken care of! The scheme was one of the highlights of my year. I'm looking forward to further development and getting the system ready for public release.”
As one might expect, the opportunities for synergy arising from initiatives such as the Summer of eResearch are immense.
Many of the solutions designed by the students and their mentors have a wide range of applicability, both within the same and other fields.
A good example of this is the Data analysis of drug discovery datasets project. Software allowing this kind of visualisation of data could also, for example, be applied (after suitable modification) to projects such as the Model Ecosystem. Another, the Model Ecosystem, has much overlap with the Invader Genetics project, and with some time and investment, a more generally useful system might well emerge from combining both approaches.
The potential for future scientific collaborations are the exciting implication of such discoveries!
There is a range of problems out there, from large, data intensive computational problems to smaller challenges that simply need some specific tools (eg. visualisation of a dataset). While scientists don’t have the time (or, often, skills) to solve these problems, they are perfect fodder for young, keen computer scientists: for example, over the period of only a few months, a small number of students (some still undergraduate) were able to move forward some cutting edge science.
The general consensus is that NZ’s first Summer of eResearch was a great success: both students and mentors benefitted greatly from the projects, and there is scope for far more development with the appropriate mentorship and funding.
We look forward immensely to seeing what next year’s Summer of eResearch projects achieve!