The Organization for Economic Cooperation and Development (OECD) has been awarded the inaugural University Press Redux Sustainability Award for its SDG Pathfinder website, an innovative tool that links all its content to relevant goals from the United Nations Sustainable Development Goals (SDGs). The goal of the site is to support decision-making, and to encourage knowledge sharing.
Using UNSILO technology, the Pathfinder combines machine learning and human curation to analyze policy documents and data published by the OECD and its partner IGOs (United Nations, World Trade Organization, Commonwealth Publications, Nordic Co-operation, International Telecommunication Union). The SDG Pathfinder is fully open-access platform providing authoritative analysis and data on the Sustainable Development Goals (SDGs).
The system identifies any documents that correspond with one or more of the seventeen UN SDGs. Using UNSILO Classify, the OECD Digital Content and Innovation team was able to tag many thousands of documents for these 17 categories without the need to employ a team of taxonomists.
Users can explore all the concepts that comprise each of the SDGs, and can look at relevant documents containing any of those concepts.
In addition, via an innovative topic map, users can see at a glance which documents in the collection correspond to specific goals. They can also select goals by country or by publisher. Users can create a reading list to explore the goals in more detail.
In the award citation, the SDG Pathfinder was praised for its “innovative, collaborative, and thematic approach … a great way to facilitate access to research just where it is needed.”
OECD built its first implementation of the SDG Pathfinder in 2019, and it has been in continuous development ever since. UNSILO automatically identifies concepts both from OECD publications and from the content of the other five partners who publish on the OECD’s iLibrary platform.
How Classify identifies sustainability concepts
UNSILO Classify enables teams to use the power of AI to build subject collections using an AI training model. First, the team identifies relevant concepts and core articles and documents, following which the machine finds other content using the concepts identified. Part of the challenge of the Sustainable Development Goals is that the 17 goals are very broad (for example “life under water”, or “global health”), so each goal was divided by OECD Digital Content & Innovation into a number of smaller concepts, forming a cluster which together represented the full goal. For example, the goal “good health and well-being” incorporates concepts such as “diabetes”, “mental health”, and “reproductive health”. UNSILO Classify has been designed to work with clusters of concepts; it identifies over 100 concepts for each document it indexes. Over 650 core topics have been identified for the SDGs, as well as 4,000 topics indicating their scope and inter-relations.
For more details, users can follow an interactive guided tour on the SDG Pathfinder website.