In a world where timing matters, opportunities to deliver research impact can be seized in a fraction of the time.
— Johan Boshoff, Manager, CASI — Computation, Analytics, Software and Informatics data transformation hub
Translating research into outcomes that make a difference on the ground is the driving force behind CASI, UNE’s Computation, Analytics, Software & Informatics data transformation hub.
Software development is one of the fastest and most efficient ways to deliver research outcomes for end users. Solutions can be delivered in months by small teams and at modest cost. Compared with developing vaccines, which cost $500m to $1.2b and take more than 10 years, thousands of applications could be developed and placed in the hands of users that will benefit from the underlying research. This can enable a culture that is not adverse to risk, as ideas that are not viable can be determined and terminated early and at low cost. In a world where timing matters, opportunities to deliver research impact can be seized in a fraction of the time.
One of the key challenges in application development is discovering what features end users want. In a significant number of cases, end users don’t know themselves until they have a chance to visualise and play with a prototype, especially if it is a new and novel concept. Using an agile software methodology enables teams of software engineers to build concepts in short development cycles and put that in front of end users for feedback. Feedback enables the engineers to further build on the idea if the feedback is positive, but also to change or even abandon concepts if negative feedback is received. It is important to collect feedback in multiple ways, as users will not always be honest if asked. Observing how users interact with applications provides valuable insights in areas that may not be well understood or are difficult to interact with.
Using an agile methodology does create tensions with funding bodies that are looking for pre-determined outcomes in the projects they fund. It is not possible to predict the functionality end users may find valuable at the start of a project. Most agile projects have a fixed budget and a variable scope, so key to the success of this methodology is regular interaction with end users and stakeholders to prioritise important functionality.
Application development provides a great vehicle for delivering multi-disciplinary research leveraging the same data. A great example of this is the ASKBILL application that CASI developed for the Sheep CRC. ASKBILL is web-based software that provides timely and accurate predictions of sheep wellbeing and productivity using weather, stock and pasture information to sheep producers across Australia.
Weather is the biggest driver of risk and productivity on farm. To enable accurate modelling in ASKBILL, an extensive weather database was assembled. Weather observations and forecasts are available on a 5×5 km grid scale for the whole of Australia and major offshore islands. For each grid point, 32 years of historic observations are combined with a 5 day deterministic forecast and 30 x 180 day probabilistic forecasts. Observations and forecasts are ingested multiple times per day and up to 1.5 billion updates to the database are done per day to provide the latest forecasts used for modelling. The accuracy is further enhanced by using on-farm weather observations. This provides a personalised weather forecasting service for farmers that is not based on the closest town 50 km away, but at most 2.5 km from the source.
The fine grained weather forecasts are used to calculate daily national risk maps for flystrike and mortality to off shear sheep and newly born lambs due to cold events.
Delivering an application is a first step in ensuring longer-term adoption. A significant number of great ideas fail as potential users are not aware of the opportunities that a new application delivers. Fortunately, social media provides a platform to raise awareness to prospective audiences. Longer-term support is key to the success that is often overlooked. Almost all of the funding is targeted at the initial development, but very little consideration is given to what happens afterwards. It is like putting together a custom built car that will exactly meet the requirements of its potential users and then not ensure that there is fuel available to drive it.