The programme will sponsor PhD and Postdoctoral students to conduct research projects in the following proposed areas:
- Interpretation of weather forecasting and observation data;
- Development of adaptive learning tools to improve processes which interact with the environment (such as wastewater discharges from industrial activities);
- Analysis of the causes of drift in data streams which may be due to natural reasons (e.g. climate change causing temperature to drift, tectonic change to base level in water levels) or data quality issues (e.g. sensor recalibration or cleaning issues);
- Prediction of extreme events in rivers, lakes and climates;
- Development of techniques to separate and study anomalies in climate and environmental data;
- The automatic extraction and explanation of information from video streams (such as the detection of and identification of pest species in DoC surveillance footage);
- Tools to accurately represent predictive uncertainty (such as determining the probability of misidentifying species in video footage).
Each project will be led by a Data Scientist, with a designated Environmental Science expert (based on the case study used to develop the tool).
A significant amount of Data Science involves finding, accessing and conducting quality assurance on data sources. To support the research projects, the TAIAO platform will provide a catalog to manage and share data sources, transformations, models and visualisations or results. This will include information on data owners, licensing and code to easily connect to the data.
Over the seven year programme, the platform will be extended to include infrastructure to store / cache data sources, run transformation and models and host results. The TAIAO Platform will be Open Source and available to the wider environmental sector beyond the research projects.