What are the challenges and benefits of open data?
Professor Guo Heading, President of CODATA, introduced the event by defining the meaning of open data and its importance for the progress of science.
“Open data is data for anyone’s use, for any purpose and at no cost.”
He addressed the crucial role of open data for interdisciplinary research and for the acceleration of the scientific progress.
Professor Geoffrey Boulton’s presentation, ‘Open Data for Open Science’, demonstrated how, historically, open data is not an unknown concept. In 1665 Henry (Heinrich) Oldenburg, as 1st Secretary of the Royal Society, edited and published the Philosophical Transactions, the first journal including only scientific material. Letters submitted could not be in Latin and had to present the concept as well as the evidence, thus, ‘open data’.
Scientific publications now often omit the original data and method, leading to low reproducibility of the published content, transforming science into a ‘black box’.
In times of crisis, the practice of data sharing can, however, be invaluable. During a gastro-intestinal e-coli infection in Hamburg, the analysis and release of the genome under an open data license helped to identify the possible source more rapidly.
Using open data effectively
For open data to be used effectively it needs to be accessible, intelligible, assessable and re-usable.
To achieve such better practice, everybody involved in the publishing process, including scientists, universities, funders and publishers, has to take responsibility for their actions. These include an adjustment of the incentives and promotion criteria for scientists, revised strategies and management processes for handling publications and data sets and the acceptance of different scientific outputs.
A panel discussion chaired by Professor Kari Raivio included responses from Professor Michael Clegg, Dr. Raghavendra Gadagkar and Dr Eva Alisic, along with the wider audience. Although the evening was primarily about open data, the direct link to open access and publication practice became apparent during the discussion.
Ideas and issues raised in the discussion included:
- The role of the funding community, with its ability to enforce change, in leading enhanced data sharing.
- Sharing small data sets via publications’ supplementary information
- A related difficulty in the implementation of open data is the direct correlation of a journal’s impact factor to its subscription price – possibly associating open data with low quality.
- The need for long term data storage and accessibility (both at least for the next generation) and the associated costs of storage and human resources.
It seems that the understanding of data ownership needs to be revised and agreed on by all parties involved. A more Maori-like understating of ownership would concentrate on the responsibilities of ownership rather than on the rights and commercial effects.