Project 21: Monitoring and predicting protection risks in inaccessible areas
Challenge
In Niger, humanitarians face increasing barriers to accessing displaced and host communities in certain conflict-affected areas – making it challenging to collect data on potential risks (for instance, conflict or food insecurity) to support the design and implementation of programming to safeguard these communities.
Solution
Developing a machine-learning model that analyses publicly available data to assess and predict protection risks in inaccessible areas. This innovation will enhance the capacity of UNHCR and its partners to monitor and respond to the needs of vulnerable populations in regions where traditional data collection is not feasible.
Impact
More effective protection monitoring – thanks to reliable data and analysis on potential risks in inaccessible areas – will support the design of tailored humanitarian interventions and advocacy, ultimately improving the lives of vulnerable communities and furthering UNHCR's protection mandate.
Other information
Project 21 (P21) transforms fragmented Household and Key Informant assessments into a unified Protection & Early Warning System that combines an AHP-weighted composite Protection Score/Index, hybrid spatial-temporal machine learning (LightGBM) and one-month-ahead risk projections. Find out more on the project website.