Sun C1, Guo S1
1Sydney Informatics Hub, Core Research Facilities, The University Of Sydney, Camperdown, Australia
Biography:
Dr Chao Sun is the Data Science Group Lead at the Sydney Informatics Hub, Core Research Facilities, University of Sydney, overseeing two development teams that support national research platform projects. With a background in electronics engineering and machine learning (PhD), Chao has spent over a decade collaborating with domain experts across disciplines. Chao began in the Faculty of Arts and Social Sciences, supporting humanities and social science researchers, then expanded to more faculties after joining the Sydney Informatics Hub in 2019. Since 2022, Chao has led development of the text analytics suite for the Language Data Commons of Australia (LDaCA), an ARDC–dHASS partnership, and since 2024 has contributed to the Australian Plant Phenomics Network (APPN).
Abstract:
Shared research infrastructure cannot serve every analytic need for all researchers and should not have to. During the early development of Wordflow, the open-source web-based text analytics tool of the Language Data Commons of Australia, a critical question for us was whether it is possible to serve a growing number of users sustainably on the project's very limited budget for hosting, and beyond the project's lifespan. The answer to this, and to other common concerns such as trusted research environments, processing protected data and very large datasets, was to offer various deployment options to suit diverse needs and concerns. We made the same codebase and final product deployable in four different ways, each suiting a different use case whilst maintaining a uniform experience.
Australian and New Zealand researchers can meet the tool on the Australian Research Data Commons (ARDC) Binder service through institutional sign-in, with nothing to install. Therefore, it's ideal for trials, classrooms and training sessions, when the users are not planning to resume their work at a later stage.
For those who need to resume their work from anywhere, e.g. on hot-desks or IT-restricted computers, we provide a persistent web host on Nectar that supports multiple users; this also means that the limited computing resources are shared among concurrent users, and the host may not outlast the project.
When the work gets more serious, or concerns about data privacy emerge, the bundled Windows and macOS desktop applications become the solution. With a locally installed app, the user's sensitive data never leaves their machine, and computing power scales with their own hardware, whilst the interface and user experience remain identical, online or offline.
And for users who are comfortable with programming and wanting full control, a one-line Python Package Index (PyPI) command opens a local host on their own computer.
The various technical solutions are not the point; the sustainability benefits are. When the heavy and privacy-bound workloads shift onto researchers' own devices, the shared national infrastructure (Nectar and ARDC BinderHub) can serve more users with the same resources. For the development team, the only extra cost is a one-off effort to build an automated release pipeline that produces all four channels. On top of standard English support, we enabled multilingual support, including Chinese, Japanese, Korean and more, which carries the tool to international communities without drawing on nationally funded hosting.
The tool is free to inspect, reuse and extend under open licensing, echoing the FAIR and CARE principles ARDC promotes. We will share what we built, the lessons we learnt, and how this model is perceived by others.
