Ensuring responsible dissemination of scientific data*

Discussions at the Global Biodiversity Informatics Conference centred on this question: How can the new frontiers in biodiversity science, computing technology and social networks combine to help answer some of the critical questions essential for better management of the planet's biological resources?

A recent example of increasing societal demand for scientific data was observed during and after the Fukushima Daichii nuclear disaster following the Great Tohuku Japan earthquake and tsunami.

Delegates at the Global Biodiversity Informatics Conference held in Copenhagen in July 2012.

View larger version

By Wim Hugo, Chief Data and Information Officer, SAEON

The scientific community has traditionally been the main audience of scientific data activities of data centres and data service providers.

However, expectations from society to access scientific data are dramatically increasing, particularly for data linked to environment and disaster issues.

A recent example of such societal demand for scientific data was observed during and after the Fukushima Daichii nuclear disaster following the Great Tohuku Japan earthquake and tsunami. Low dose radioactivity contamination data attracted attention and were scrutinised by the general public. Unfortunately, raw and complex data were exposed to an unprepared public without proper scientific explanations - leading to confusion in the Japanese society.

This illustrated how the public trust in science can be seriously compromised when no proper mechanism is available to release scientific data to the public. New measures are required to promote the appropriate sharing of scientific data with the general public, as well as to increase the understanding of scientific data and information. These include education and outreach activities involving scientists, data providers and data managers.

Improving public trust in science

With this in mind, CoDATA Taipei is considering a special session to study the impact of freely available scientific data on society.

Expectations from society to access scientific data are dramatically increasing, particularly for data linked to environment and disaster issues.

This topic was raised again in a recent workshop that formed part of the Global Biodiversity Informatics Conference in Copenhagen (July, 2012). The workgroup discussing the topic concluded that frameworks are needed to guide the presentation of scientific data in general, and biodiversity data in particular, and that responsible presentation of the data required additional context, some of which can be produced automatically.

The topic also links to the on-going discussion in global data infrastructures such as the Group on Earth Observations (GEO) and the International Council for Science (ICSU) World Data System about consistent quality and usability assessments of data that are freely available in these infrastructures.

Proviso

There is one proviso: it is vital that one should not limit access for fear of misuse or misinterpretation, and a danger remains that free and open access to data can be constrained by self-appointed gatekeepers. For this reason, properly contextualised access is the ideal solution. Such proper contextualisation entails two main aspects: minimum quality information about the data, and minimum context, which in properly described data sets can be as simple as the abstract and caveats on application of data which are routinely included in meta-data.

SAEON will shortly embark on two technology projects that support this notion:

The definition of an environmental indicator exchange format. By defining this format, the probability that an indicator will be accompanied by a proper context and comparative or normative data is increased. If one looks at the quoted example of radiation levels, the proper context would be to at least also report the bounds of acceptable exposure in respect of time and intensity.
Development of Quality and Usable meta-data: GEO is in an advanced stage of finalising a quality indicator that combines assessments of syntactic and schematic quality (for example, automatable tests that confirm the accuracy, precision, structure and format of data), usability (based on assessments from users, some of which may be designated as experts and contribute more weight), and usage statistics. This will also be an obvious improvement compared to a situation of completely unqualified access.