Training on Machine Learning and GIS for Land Cover Estimation Mission to Vanuatu Bureau of Statistics

The Economic and Social Commission for Asia and the Pacific (ESCAP) is implementing a project supporting national statistical offices in Asia and the Pacific to leverage innovative data sources, tools and methods for the streamlined production and use of better, more timely data for official statistics: “The 2030 Data Decade – Strengthening the institutional capacity of national statistical offices in Asia and the Pacific to use innovative, new and big data sources for official statistics in support of the 2030 Agenda for Sustainable Development”, which is funded through the 2030 Agenda for Sustainable Development sub-fund of the UN Peace and Development Trust Fund (UNPDF).

Following a request received by the Vanuatu Bureau of Statistics (VBoS), as part of this project, ESCAP is providing technical support and guidance to Vanuatu for the integration of big data solutions into national development strategies for official statistics. Through initial engagement with VBoS, Land Accounts were identified as a key building block for addressing many identified needs relating to climate, the environment and agriculture. Producing land accounts will require land cover and, ideally, land use data, for two or more points in time. For Land cover data, it is proposed to build on the work that VBoS has done in partnership with Radiant Earth, subject to review of the sources, methods and code. This utilises the Forest Inventory as a ground truth dataset, which was made available to VBoS by the Department of Forestry. 

To support this work, further technical training is needed for VBoS staff and other Vanuatu government ministries. Training on Data Science and Machine Learning (ML) methods will be especially important to allow staff to utilise the concepts, methods and tools developed by Radiant Earth in this project.