EFSA taps into the crowd: data mining, geospatial mapping

Can you create an algorithm to extract meaningful data from a wide range of scientific literature? Do you have ideas for depicting the uncertainties linked to data? Your idea could win you a cash prize and help us to provide scientific advice that is essential for food safety in Europe.

These are the first two crowdsourcing “challenges” that EFSA is posing to people scientists and other potential “solvers” of these questions.

Untapped reservoirs

Speaking about the challenges, EFSA’s Executive Director Bernhard Url said: “There are vast untapped reservoirs of scientific knowledge and technical skills that could help us to do our risk assessments and other scientific work.

“We’re looking for up-and-coming researchers, retired professors, skilled technicians working in other fields or passionate thinkers with innovative ideas. And they could be anywhere in the world, not just in Europe.

“Crowdsourcing offers a way to reach that potential expertise and harness it productively so that it brings benefits to society and to the solvers themselves.”

Visualising scientific uncertainty

Didier Verloo, who is in charge of the team running EFSA’s crowdsourcing project, said: “The first challenge launched today is about visualising the uncertainties – what we call the limitations in available knowledge – that affect the range and the likelihood of possible answers we give in our assessments.

“Risk maps are frequently used to visualise geospatial differences, but they normally don’t show the uncertainty of the underlying data. So, for US $5,000, we are looking for a fresh and innovative idea of how to visualise uncertainty in maps.

“Besides helping assessors to contextualise their conclusions this will also help risk managers to quickly and more easily take account of uncertainties and factor them into their decision-making.”

Data mining of scientific literature reviews

The second challenge, with a prize of US$30,000, will be launched on 1 December 2017. Dr Verloo said: “Systematic review is a cornerstone of evidence-based research. But it poses challenges due to the growth in the volume evidence available and the greater demands on resources, especially the human expertise and time, needed to extract the relevant information.

“Automation of this ‘data extraction’ step would save time and money. Some techniques – natural language processing, machine learning, data mining and text mining – have started to help in screening the title and abstract in systematic reviews. But, we want to go beyond this step and automate the entire data extraction step.

“So we are looking for a software genius or computer nerd who can build information extraction algorithms to identify data elements from a full text scientific article.”

 Risk assessors of the future

Dr Url added: “I’m optimistic that more initiatives like these can help us keep pace with society’s demands for scientific advice. They will also create new opportunities, both to improve the quality of our advice and to find potential risk assessors for the future.”

Questions & answers

Who can apply?

Anyone can apply after registering as a solver on the InnoCentive platform. It is quick and easy, requiring only five simple steps.

Why are the prizes in dollars?

Solvers registered on the InnoCentive platform come from across the world so the prizes are in U.S. dollars.

How is EFSA involved in crowdsourcing?

In 2015 we launched a project on “Crowdsourcing: Engaging communities effectively in food and feed risk assessment” to explore the risks and opportunities in applying crowdsourcing and citizen science as an innovative way to collect and process data and increase the openness of EFSA.

Why is EFSA interested in crowdsourcing?

Several factors making crowdsourcing appealing for us: engaging the public and stakeholders in the process of scientific assessment, widening our evidence base, promoting data sharing and re-use of data, better use of expertise, encouraging public enthusiasm and support for science, and increasing transparency and trust in science.

At what stage is the project now?

Our project partner ADAS is launching two challenges on behalf of EFSA on the InnoCentive platform. We also have an open call for tenders (closing on 30 November) to pilot further crowdsourcing initiatives.

Is EFSA part of other European crowdsourcing initiatives?

EFSA joined (as an observer) the Citizen Science Cost Action where we participate on activities related to scientific quality, standardisation and interoperability of data.

Media contacts