by BRADLEY SMITH and PETER BENTLEY
Australian universities generate and report a huge amount of data to the Government and its agencies. The Australian Research Council and National Health and Medical Research Council also collect detailed data through their grant programmes. Generating and reporting institutional data is expensive and time consuming, particularly for grant applications where there are significant opportunity costs for staff through low success rates.
Not all data reported is provided back to universities or the public in a complete, timely or accessible manner. The asymmetry between the data provided and the ability to analyse it is frustrating.
In this context, it was pleasing to see the ARC release their interactive data visualisation tool on the 18th of November, 2020. This continues progress by the ARC and Australian Government in improving higher education data accessibility and utility (e.g. finance, students and staff).
This tool provides visualisations of Discovery and Linkage projects in the National Competitive Gants Program from 2001 using Microsoft Power Bi.
Most data can be explored at institutional, university group, state/territory and national level, by programme, year of commencement, STEM/HASS, four-digit Field of Research (FoR), success rates, lead chief investigator distribution and international collaborations (by country, not institution).
The tool supplements data available through the ARC’s Grants Search function which provides web search and downloads by a public application programming interface (API) and Excel.
So how useful is it? The new tool effectively brings together data currently spread across multiple spreadsheets and reports into a visual pleasing and analytical format. Presentation is crisp with easy to use dropdown menus and filters. The visualisations will be useful for time-poor managers interested in a quick institutional overview, as well as those looking to interrogate the detail.
The dashboards can be quite cluttered and are best viewed in full screen mode. Some users may find too many things going on, but each chart can be viewed separately in “focus mode”.
The interface allows users to quickly test preconceived ideas with detailed data. How important is China as a collaborator on ARC-funded projects in HASS versus STEM? China is the 4th most frequent collaborator in HASS (433 projects since 2002) and 5th in STEM (1,611 projects).
Disgruntled researchers and disciplinary groups will interrogate success rates by field with some alacrity. Are success rates higher in econometrics than marketing? Yes, 37 per cent versus 14 per cent (but correlation isn’t causation!).
For us, there are three non-trivial shortcomings.
First, the export function from Power Bi (e.g. to Excel or pdf) is disabled. This greatly limits the ability to share insights and data in documents and on-line. Given that the dashboards draw upon public data, we hope the ARC enable this function promptly.
Secondly, there is no capacity to analyse trends by gender and academic level. The sector shouldn’t have to rely on individual researchers analysing the API sourced data at their own initiative to generate time series tables for circulation on, e.g. Twitter (see @GaetanBurgio – 14/11/2020).
Thirdly, it would be useful to have collaboration data at an institutional level. For national policy, this would improve understanding of the intensity of domestic collaboration in particular fields, such as between regional and research-intensive metropolitan institutions. Internationally, it would help identify important collaborators, such as the aforementioned Chinese HASS projects with universities in Hong Kong and mainland China.
We understand that the ARC intends to increase the scope of the tool and we look forward to those developments.
Bradley Smith is senior research policy advisor at James Cook University
Peter Bentley is the Policy Advisor for the Innovative Research Universities (IRU)