Although predictive analytics is not a new field, its application in humanitarian response has only just begun. Predictive analytics is the analysis of current and historical data to anticipate an event or some characteristic of an event (its likelihood, severity, magnitude, or duration). In the context of humanitarian response, predictive analytics uses modelling methodologies aimed at anticipating humanitarian needs arising from shocks, including:
- Weather and climate models predicting the occurrence and impacts of sudden and slow onset hazards such as droughts, floods, storms.
- Epidemic models used to better understand the main risk factors leading to major outbreaks or to project epidemic trends.
- Analysis aimed at projecting the evolution of ongoing or upcoming humanitarian crises based on current and historical data.
- Other methods that assess the risk and the socio-economic impacts of disasters or other shocks.
The Centre’s Focus
The Centre’s predictive analytics team is focused on increasing the trust and adoption of models to support improved decision making by humanitarian actors. We do this through the following activities:
Technical support
- Assess available models and forecasts, and conduct historical analysis.
- Design trigger mechanisms for use in anticipatory action frameworks.
- Act as a technical translator to support decision makers.
Responsible analytics
- Facilitate the Centre’s Peer Review Framework for Predictive Analytics in Humanitarian Response.
- Promote data responsibility with how data is shared and used for model development.
- Support decision makers with understanding uncertainty in projections.
Capacity building
- Provide documentation, tools and training on predictive analytics.
- Offer model deep dives and events on predictive analytics in the sector.
- Create a community of technical actors working on this topic.
Anticipatory Action
One of the main applications of predictive analytics is anticipatory action. OCHA facilitates the development of collective anticipatory action frameworks to enable humanitarian organizations to get ahead of shocks and mitigate their impact.
The Centre supports anticipatory action by providing technical support for the development of triggers, engaging with technical experts and scientists, and by bridging the gap between forecast producers and humanitarian decision makers. The details of our focus countries and technical documentation are available here.
Peer Review Framework
The Centre has established a Peer Review Framework for Predictive Analytics in Humanitarian Response. Through peer review, the Centre seeks to ensure models can be understood and trusted by all stakeholders in a humanitarian operation. The Framework consists of six steps:
Partners interested in submitting a model for peer review are asked to submit a request through this form.
The Centre is looking for independent experts who can support the review of predictive models submitted by our partners. We invite experts to submit an application to become a Reviewer in the technical or ethical domain.
A Catalogue of Predictive Models
The number of models being developed to inform humanitarian action is growing rapidly. To help keep track, we have created a catalogue of predictive models with basic information on ‘who is doing what, where and when’.
If your organization is developing or using a predictive model, please share information in this Google form.
For any other questions or further information, please contact centrehumdata@un.org.