By Natalie Grover and Emmanuel Letouzé, November 15, 2018
A session on “The Open Algorithms (OPAL) Project: What has been achieved in Senegal and Colombia and how can it scale?” took place October 24 2018 at the 2nd UN World Data Forum held in Dubai on October 22-24, 18 months after the official start of the project announced at the 1st World Data Forum in Cape Town in January 2017 (picture below).
The session had two key objectives:
present the main advances and lessons of the pilots deployed in Senegal and Colombia with the countries' National Statistical Offices and leading telecom operators Orange-Sonatel and Telefonica Colombia since early 2017 with 1.5 M EUR in core funding from the French Development Agency (AFD) as well as complementary funding from The World Bank, The Global Partnership Sustainable Development Data (GPSDD) and the Sustainable Development Solutions Development Network (SDSN).
discuss opportunities and requirements for scaling OPAL, both to other countries and industries in the next few months and years, by developing and diffusing technological systems and governance standards for the ethical and scalable use of ‘big data’ sources collected by private companies for public good purposes.
The session gathered about 100 attendees and speakers from some of OPAL’s main partners since its start: Dr Claire Melamed, CEO of the GPSDD, M. Babacar Ndir, Head of the National Statistical Office of Senegal (ANSD), Ms. Sandra Liliana Moreno, Head of Geostatistics at Colombia’s National Statistical Office (DANE), Dr Seydina Ousmane Sene from the Senegalese NGO IPAR, Dr Pedro de Alarcón, Head of Big Data for Social Good at Telefonica, M. Nicolas de Cordes, VP of Marketing Anticipation at Orange Group and member of the Executive Committee of OPAL, M. William Hoffman, Head of Data Policy at the World Data Forum and member of the Executive Committee of OPAL, and Dr Emmanuel Letouzé, OPAL Program Director and Director of Data-Pop Alliance, moderated by Ms Natalie Grover, OPAL’s Global Program Manager.
Most of the exchanges and questions focused on OPAL’s potential to allow NSO and others to ‘leverage’ private sector data systematically for SDG monitoring and promotion and other policy uses, as well as ethical considerations regarding statistical and political representation of data subjects, among others.
During the opening part of the session, Pedro de Alarcón, Nicolas de Cordes, and Emmanuel Letouzé demonstrated live, for the first time ever, how the platform worked ‘end-to-end’, by allowing users with the appropriate credentials to select and run an open algorithms directly on the servers of the data partner companies to compute and extract key statistical indicators, such as, currently, population density estimates, from cell phone activity.
Later, M. Ndir and Ms Moreno explained their current involvement in the project and expectations that in the medium term indicators coming out of OPAL could feed into the standard statistical production process, while incentivizing other use cases and helping raise awareness and capacity levels within their organizations. Dr Sene described his experience as one of OPAL’s initial Friendly User Testers (FUTs) on a project utilizing population density estimates for rural development, and as a member of OPAL’s CODE, for Council for the Orientation for the Development and Ethics, the project’s core governance and oversight body in Senegal.
Such is the ‘gist’, DNA, and revolutionary potential of OPAL: unlocking the tremendous social value contained in these sensitive datasets, without encroaching on users’ privacy, safety, and agency, to provide public, private and civil society organizations and regular citizens, a much better, fine-grained, picture of human reality to make better-informed decisions, while allowing individuals to have a greater say in how and by whom data about themselves are used. As such, building on several years of work by its founders and others, OPAL is seen a key milestone of a vision where data will be at the heart of 21st Century social contracts and systems, in the context and in support of the Sustainable Development Goals.
Visuals displayed in the room depicted OPAL’s 2 main components and ecosystem reflecting and contributing to these objectives:
A Technology track, with a state-of-the-art privacy-preserving open source platform and open algorithms, developed by Imperial College London, MIT Media Lab, Orange Labs, Telefónica, and communities of local developers. With OPAL’s “question-and-answer” approach to data access and analysis, computation of key indicators is “pushed out” through certified open algorithms running directly on pseudonymized data that remain on the servers of the partner companies, behind their firewalls. As a result, no sensitive data ever leave the servers to be exposed to theft and misuse, and only aggregated statistics are made available to selected users.
A Governance track, with a local CODE in each country, participatory design and testing with local users, and capacity building activities, led by Data-Pop Alliance and the World Economic Forum, and local partners. This aims to ensure that use cases are relevant to local needs and constraints, respectful of local ethical norms and regulations, and that capacities, incentives, and connections are strengthened within local ecosystems. Building first generations data governance standards centered on empowerment of data subjects, is a core objective of OPAL.
Last, OPAL’s future plans were discussed; starting early 2019, it will be moving into a “Beta phase: with the objective of strengthening its foundations in both pilot countries and expand to 1 or more countries by end of 2020, and up to 10 countries and 2 industries by 2025. Current countries of considerations for future expansion include Kenya, Ivory Coast, Sierra Leone, Mexico, Guatemala, Haiti, Tunisia, Jordan, and Moldova, notably.
Additional information will be provided in the next few weeks; in the meantime, questions and suggestions can be sent to firstname.lastname@example.org.