The World Bank has partnered with the Stockholm Environment Institute (Latin America Center) to develop a text mining methodology that uncovers how investments made by the Bank are supporting the achievement of the 2030 Agenda and the Sustainable Development Goals (SDGs).
Since 2019, the Bank and SEI have tested and developed the methodology, which allows data-driven analysis using automated software to match each project to the SDG targets it is contributing to. This type of mapping will be applied to the WB portfolio for the FY20 Sustainable Development Bond Impact Report. Projects for FY18 and FY19 will also be analyzed, and this mapping will be published on the WB website. In addition, the project will result in a working paper describing the methodology developed.
The methodology consists of two phases. The first one is a quantitative content analysis approach that involves extracting words automatically from the text to analyze them statistically. The second phase refers to creating coding rules looking for specific expressions and sentence structures – this allows the identification of concepts and meanings in the text. We designed this coding to find SDG-related topics, based on the text from the WB projects and the SDG description. The identified codes are then statistically analyzed to deepen the insights and find correlations between the WB projects and the SDGs.
Both phases intend to jointly provide a better picture of the SDGs being most impacted by the reviewed WB projects. An ambitious undertaking would be to create a coding protocol to analyze all WB projects automatically, which could have applicability to many other sectors of development.
As an important outcome of the project is fair to say that SEI inputs are very well recognized in the latest World Bank Report. Charts created by the SDG team are present on pages 24-25 and 138-139.
Design and development by Soapbox.