CrowdVsCovid : getting the right data to the right people at the right time
Our Challenge : Why it Matters
Solution : Crowdsourcing projects validating scientific articles and social media posts
A snap shot of the platform showing the tasks
Citizens can help scientists and policy makers to find relevant information about policy intervention
How the citizens can contribute in selecting the most relevant article
How the citizens can add additional attributes to the relevant article
Citizens can help scientists extract useful data from social media
Fully functional demo for analysis of tweet images
Demo showing detailed questions that participants must answer
Scientific literature key data extraction process diagream
Social media data quality enrichment process diagram
Social media AI data filtering pipeline
Social media AI data filtering tools
Steps for mapping data resulting from social media crowdsourcing
Resulting European map based on social media data
Global map based on social media data
What problem are we trying to solve?
CrowdVsCovid is a team of citizens and scientists from research institutions in France, Italy, Spain, Switzerland and the UK, keen to provide policy makers with relevant and actionable information on a range of Covid-related issues, as quickly and reliably as possible.
At both the regional, national and European levels, policy makers urgently need the best scientific information in order to make informed policy decisions. When should schools be re-opened? What proportion of the population is wearing masks in public? How is confinement affecting the mental health of different age groups?
Some scientists on our team have been asked by their national authorities to provide answers to such questions. So they brought their teams to the hackathon, to work together and with motivated hackers, to find better solutions to this problem.
Useful data can come from various sources: scientific articles and preprints, 'grey literature' reports published by a range of organisations, social media on different platforms and information scraped from a variety of websites. But the quantity of information involved is huge and growing rapidly, and the quality varies widely.
AI technologies can help to filter the most relevant results out of all this data. But the amounts of information are often still overwhelming, and requires human intelligence to further extract the most essential facts for making policy decisions.
What's new about the CrowdVsCovid solution?
To tackle this challenge we are combining automated filtering using AI algorithms with crowdsourced refinement of the resulting data, in a way that will enable motivated citizens to participate in the process, building on their collective intelligence to produce data that informs policy.
In other words, we aim to combine the speed of machines with the intelligence of people. And we aim to do this in a novel way that is financially sustainable, can scale to other urgent problems well beyond the immediate Covid crisis, and provides meaningful educational rewards to citizens.
What have we achieved during the hackathon?
Over the weekend, we built on some preliminary results obtained at a previous online hackathon called Versus Virus that took place three weeks ago. We developed a full pipeline model, going from the raw data input to policy advice output.
We also made two fully functional demos, one of which we were able to test during EUvsVirus, by inviting all the hackathon participants to try it out.
The first demo focuses on reviewing scientific articles, and extracting key information based on specific policy concerns. For example, as European countries open up schools, policy makers want to know what the experiences have been in countries that have already opened schools or did not fully lock down, as well as experiences from other public health crises that may be relevant to this issue.
The second demo uses social media data from twitter. Especially during a lockdown, where standard survey techniques become impossible, analysis of social media provides useful insights, for example by analysing the types of masks that people are wearing in public places, or the sentiments of people from different regions during lockdown.
For this demo, we also created a mapping tool, so that the data resulting from the crowdsourcing could be mapped onto European countries or regions, to reveal significant variations across the continent more clearly.
Finally ran discussion sessions with members of related hackathon projects, with mentors and with several external experts, including a researcher at a national statistical office, and the co-founder of a European crowdsourcing platform, called WeMakeIt, in order to build a sustainable business plan for deploying the AI and crowdsourcing solutions we were able to demo.
In brief, the business model is based on organisations such as public authorities sponsoring challenges in which motivated citizens crowdsource large amounts of pre-filtered data. Thus, the citizens provide in-kind contributions rather than cash, and in turn they earn certificates and credits from the Universities involved, to reflect the effort they invest in carefully reading technical articles or analysing other forms of data. The Universities manage the AI-to-crowdsourcing workflow and produce the resulting policy advice, thanks to the challenge sponsorship.
How could CrowdVsCovid impact the crisis?
The inspiration for this team were the requests for information that two of the scientists involved had received directly from policy makers at the national level in their country. The challenge of getting the best information to these decision makers as rapidly as possible will continue even after lockdown conditions are relaxed, since there will be major public health as well as broader economic and social consequences for months and possibly years to come, often requiring new policies.
What's next for CrowdVsCovid?
Having successfully tested the crowdsourcing demos on relatively small data sets of about a hundred publications or social media images, the team is ready to scale up to much large amounts of data, and to connect AI filtering seamlessly to the crowdsourced evaluation of the most promising results. Thus, we are just a few steps away from practical and large-scale deployment, and producing actionable data for policy makers.
How could our solution be used beyond the Covid crisis?
Several of the participants plan to use the experience gained during this hackathon for an EC H2020 research and innovation project called Crowd4SDG that they are part of, which is about crowdsourcing for the sustainable development goals, with a particular focus on climate action. This new project, that will be launched in May, will be able to adapt the demos developed in this hackathon to provide valuable knowledge for policies that tackle climate-related extreme weather crises. In parallel, these participants will test a longer-term sustainable business model on the crowdfunding platform mentioned above.