Welcome to The GovLab’s #Data4COVID19 Round-Up. This temporary, weekly curation will provide you with notable updates to the #Data4COVID19 data collaborative repository and other information meant to facilitate data stewardship in the ongoing pandemic. Have an interesting novel coronavirus-related project worth sharing? Send it to us at email@example.com so we can consider it as an addition.
Highlighting Efforts of Call for Action Signatories
Two weeks ago, we at The GovLab wrote a reflection on our call for action. Written last year at the start of the pandemic, the call urged organizations to act to foster the data ecosystem needed to use data responsibly to address threats like the COVID-19 pandemic and has been signed by over 500 data leaders and practitioners across 63 countries. As part of our reflection, we urged signatories to send us explanations of their work to foster a more responsible data ecosystem.
Week by week, we intend to highlight responses to that call. This week, we focus on CoronaCheck, an initiative submitted by call signatory and EURECOM professor Paolo Papotti. As described by Dr. Papotti:
“The idea of fact checking is to verify if a given claim (sentence) is true or false, based on reliable sources or data as a reference frame. In this context, we started “CoronaCheck”
(https://coronacheck.eurecom.fr), an artificial intelligence project that aims at helping people to verify online information related to COVID-19, currently operational in seven languages. More precisely, CoronaCheck checks the reliability and accuracy of the news and statistical claims circulating online regarding the coronavirus, based on the official data sources.
CoronaCheck verifies the veracity of a statistical claim related to the COVID-19 pandemic. For example, ‘COVID-19 cases raised by 10% compared to last month in France’ or [the] ‘death rate in Europe is higher than in [the] USA.’ CoronaCheck provides an explanation for any statement found to be ‘True’ or ‘False,’ which is what makes our tool complete and reassuring to use. It is a feature that is not obvious to implement with other ‘black box’ machine learning approaches, but it comes naturally with our approach based on queries over structured data.”
CornaCheck has been used by more than 15,000 unique users over the past year and is supported by gifts and grants from Google.org and the International Fact Checking Network. Additional information on the project can be found at the Medium post here, YouTube video here, and through the technical paper here.
We encourage others to email us at datastewards [at] thegovlab.org describing what projects they’ve undertaken to support responsible, data-driven approaches to COVID-19 and how they relate to the Call for Action. Include “A Call for Action” in the subject line. If relevant, we may highlight your work here and promote it to others in our network of experts. A project’s inclusion does not indicate endorsement by The GovLab or confirmation of its success in meeting its goals.
Analyzing Open Government Data Efforts from the Early Stages of the Pandemic
Last week, The GovLab and the OECD published Open data in action: initiatives during the initial stage of the COVID-19 pandemic. This report analyzes a series of open government data initiatives launched from March through July 2020 to understand how governments initially reacted to COVID-19, what issues and approaches they prioritized, and what gaps existed in their efforts.
An analysis of 85 initiatives found that governments were active in releasing and re-using open government data and that it played an important role in communication efforts. However, these efforts often did not address economic or social needs and tended to focus more on responding to immediate needs than long-term recovery or reform.
The full report can be found here.
Analyzing the #Data4COVID19 Repository
A year since we published our call for action, The GovLab has focused more of its efforts on understanding patterns and assessing the examples it has collected. As of this week, the repository has 237 examples, spread across six continents. These include:
- East Asia and Pacific: 11 (4.6%)
- Europe and Central Asia: 47 (19.8%)
- Latin America and the Caribbean: 7 (2.9%)
- Middle East and North Africa: 1 (0.4%)
- North America: 76 (32.1%)
- South Asia: 4 (1.7%)
- Sub-Saharan Africa: 13 (5.4%)
- Global: 90 (37.9%)
These projects can be mapped onto one of six data collaborative types. First, they can be examples of public interfaces, projects in which organizations provide open access to certain data assets, enabling independent uses of the data by external parties. Second, they can be trusted intermediaries, efforts in which third-party actors support collaboration across sectors and data users from the public sector, civil society, or academia. Third, they can take the form of data pools in which data holders agree to create a unified presentation of datasets as a collection accessible by multiple parties. Fourth, they can be research and analysis partnerships, arrangements in which data holders engage directly with public-sector partners and share certain proprietary data assets to generate new knowledge with public value. Fifth, they can be prizes and challenges in which data holders makek data available to participants who compete to develop apps; answer problem statements; test hypotheses and premises; or pioneer innovative uses of data for the public interest and to provide business value. Finally, they can be intelligence generation efforts that allow organizations to internally develop data-driven analyses, tools, and other resources, and release those insights to the broader public.
Our early analysis suggests the repository is composed of the following models:
- Public Interfaces: 66 (27.1%)
- Trusted Intermediaries: 12 (4.9%)
- Data Pools: 42 (17.2%)
- Research and Analysis Partnerships: 53 (21.7%)
- Prizes and Challenges: 9 (3.7%)
- Intelligence Generation: 62 (25.4%)
While almost all data collaborative types have significant representation in the repository, we note the lack of examples of prizes and challenges. We speculate this disparity could be the result of difficulties in balancing sensitive data and the protections required to make it accessible to distributed challenge participants. Next month, as part of the COVIDaction initiative, The GovLab and University College London will release a new study to better understand the utility and impact of data prizes and challenges to help address COVID-19.
While the repository should not be interpreted as comprehensive or representative, these figures can suggest gaps in ongoing work. In the coming weeks, The GovLab will publish additional tallies of the repository. We hope you will follow along with this work.
You can also find additional resources related to data stewardship and data collaboration here. Have an interesting novel coronavirus-related project worth sharing? Send it to us at firstname.lastname@example.org so we can consider it as an addition.