Ken Kahn’s report on a selection of the presentations at the Digital Research Conference:
Citizen Science and Crowd-sourcing Biological Data — Developed a phone app that greatly improved previous citizen science projects that relied upon email with photo attachments. Most examples were reporting the health of trees by submitting photos of leaves. One used the microphone to capture sounds made by bats.
The Environmental Virtual Observatory: A user-driven cloud-implementation of environmental models and data for all — data (often real-time) combined with models (with user-settable parameters) for both local and global environmental data. Initial focus on hydrology and soil since models are more local and much easier to compute.
A Case Study in The Data Science Approach – Finding the Best Beer in Oxford — defined a data scientist as a combination of data analyst, software developer, and story teller. He scrapped ratebeer.com and other sites to collect descriptions of beers and pubs. Extracted frequency of unusual words in order to find the pub that sells beers that best matches his own tastes.
Mining and Mapping the Research Landscape — Very similar to the above but instead of beer it was matching reviewers with submitted papers or matching papers with researchers doing similar work.
Will the real data scientist please stand up? — About the Open Data Institute that helps startups, offers training and certifies sites.
Data science on the cloud — About what Microsoft (and Microsoft Research) are doing with the cloud. Azure seems like a pretty good service and they are eager to support and showcase research by academics and their data and analyses.
The Behaviour Composer: A web-based tool for authoring, exploring, and sharing agent-based models and behaviours — Ken Kahn’s talk.
The final Research Technologist Panel was about the different roles for support and collaboration by software developers with research projects. The UCL Research Software Development team was particularly interesting. In addition to funded projects they take on one project every three months (UCL researchers propose projects and one is awarded support). They claim (plausibly) that the resulting code is much more professional (correct, robust, documented, readable) than is the norm for university research projects.