Challenge: Currently, countries at high risk of humanitarian emergencies are identified jointly by UNHCR operations, regional bureaux and headquarters, but the agency lacks systematic, data-driven early warning systems on forced displacement that provide real-time alerts. We must boost UNHCR’s ability to forecast in order to build our capacity to act.

Solution: UNHCR’s Division of Emergency, Security and Supply will design a prototype data-driven early warning system combining different data sources, based on research of internal initiatives and external systems used by humanitarian agencies, academia and the private sector. The system will be piloted in several country operations.

Expected results: Improved anticipation of and response to complex emergencies, strengthened partnerships, and a role in data-driven anticipatory humanitarianism.

Challenge: The Research, Strategy and Analytics team of UNHCR’s Global Communications Service has developed a model that classifies hateful content in English targeting forcibly displaced people. This partial solution needs to be operationalized to systematize and automate data collection.

Solution: A user-friendly web application that tracks, visualizes and catalogues hateful social activities almost in real-time to facilitate analysis and alert key stakeholders via established communication channels when crises are identified.

Expected results: Improved hate speech detection and response within UNHCR in accordance with the organization’s data protection standards, allowing UNHCR and its humanitarian partners to better understand hateful activities online and prevent discrimination against displaced people.

Challenge: A lack of appropriate sampling frames makes it hard for UNHCR to get reliable data on people forced to flee from household surveys.

Solution: Three of UNHCR’s Data and Identity Management Analysis (DIMA) units teamed up to design a semiautomated delineation for enumeration areas, using satellite imagery and employing Bayesian statistical modelling techniques to estimate target population size from secondary data sources.

Expected results: Readily available sampling frames leading to reliable household surveys for country operations, reducing the complexity and costs of carrying out survey activities and improving the resulting data.

Challenge: UNHCR Afghanistan gathers a vast amount of text-based qualitative data through community-based protection monitoring, but manually processing this information is time and energy intensive, resulting in lost opportunities to use this data to inform timely interventions.

Solution: Use AI-based software, particularly with natural language processing features to facilitate qualitative analysis of data gathered through community-based protection monitoring. This software would detect key words and phrases, categorise and translate them, and produce meaningful analysis.

Expected results: Effective and efficient data analysis, resulting in better use of data to inform effective protection interventions and timely responses to protection risks.

Challenge: Refugees and asylum-seekers in Kenya have historically been excluded from national social protection systems. However, when displacement is protracted and national services are developing, humanitarian actors are expected to demonstrate why they’re not working with national systems.

Solution: Integrate refugees and asylum-seekers into Kenya’s national registry to enable local solutions and access to social services, better target assistance, and broaden stakeholder engagement to support displaced populations as well as host communities.

Expected results: Improved integration of refugees in a way that benefits displaced people and host communities, reduces response times, avoids duplication, supports local economies and livelihoods for refugees and asylum seekers.

Challenge: With shifts to digital learning, displaced communities are excluded from the benefits of these types of learning opportunities due to a lack of infrastructure, devices and know-how. UNHCR has worked to track and centralise digital learning resources but needs to more rapidly align content to relevant curricula, in partnership with Learning Equality.

Solution: UNHCR’s Global Education team will use machine learning and natural language processing to partially automate the curriculum alignment process, which will improve the relevance, adoption, and impact of digital learning resources. A competition is bringing together machine-learning practitioners to design new algorithms to support this effort, and the winning solutions will be made publicly available.

Expected results: Creative solutions governments and educators can leverage to more efficiently align digital resources, particularly during crises. Uninterrupted access to education is supported for displaced communities and others, with benefits for students, teachers, parents and communities.

Challenge: UNHCR Guatemala lacks an evidence-based methodology to prioritise communities most in need of community-based protection, affecting its ability to reach and respond to the most vulnerable populations, who tend to be widely dispersed across different contexts.

Solution: The design and construction of an evidence-based community intervention prioritisation tool that allows for a more objective, systematic, and reproducible prioritisation approach. A participatory design methodology facilitates input from the people UNHCR serves, community leaders, local authorities, and experts.

Expected results: A customizable online tool that is easy to use in different contexts, settings, and operations and that will enable UNHCR to better understand the situation of communities. This will support the prioritisation and tailoring of interventions, improving outcomes for the most vulnerable communities.

Challenge: Over 222,000 refugees live across more than 3,500 informal tented settlements in Lebanon’s Bekaa Valley. Standardised shelter criteria determine households’ eligibility for additional shelter kits, but the absence of a tent-level address system hinders this process.

Solution: An address system for informal tented settlements, using a code for each tent. Part of the code will be fixed, reflecting the tent’s location, and another part will change based on the physical condition of the tent, to better link household needs with assistance and services.

Expected results: This tent-level address system will enable UNHCR to monitor household location and conditions, enabling more efficient and accurate identification of needs and the required response, and supporting the health and safety of families in Bekaa.

Challenge: Mexico lacks an integrated data analysis framework to support joint analysis and planning by actors engaging with displaced communities. This results in inefficient resource allocation, with duplications in certain settings while many displaced people face barriers accessing basic services.

Solution: The creation of a data analysis framework to collate, study, and interpret information from the national refugee commission, the national population registry, and UNHCR. This will generate information detailing the characteristics and displacement patterns of the affected population, and highlight cases of successful local integration.

Expected results: UNHCR will better understand the needs, capacities, motivations and interests of displaced people in Mexico – information that will help decision-making processes, reduce duplication, and enable better coordination to deliver effective interventions.

Challenge: Due diligence requires that UNHCR thoroughly screen its private-sector partners and analyse potential reputational risks. The resulting labour-intensive screening reports are valid for a year, during which the donor monitoring work is more flexible and variable.

Solution: UNHCR’s BEYOND project team will use AI-based software with natural language processing (NLP) features to develop a tool that automatically feeds pertinent risk- and opportunity-related updates on partners to relevant teams, enabling ongoing risk-monitoring while facilitating strategic donor cultivation and partnership development.

Expected outcome: This tool will improve and support proactive risk monitoring and relationship work through automatic updates, as well as optimising and streamlining the research, identification, and documentation of controversies for the due diligence screening process.

Challenge: Eliminating malnutrition is integral to UNHCR’s protection mandate, but existing methods to accurately and effectively diagnose malnutrition can require skill, time, cumbersome equipment, and direct contact. This makes it hard to conduct screening at the community level.

Solution: The Severe Acute Malnutrition (SAM) Photo Diagnosis App® is an innovative tool that uses morphology to screen wasting among children. UNHCR’s public health teams propose using the app, provided by Action Against Hunger, to simplify community screening and diagnosis of acute malnutrition among refugees. It is the first time this app is deployed in refugee-contexts.

Expected results: Faster, easier, real-time, and systematic diagnoses of malnutrition, which can be carried out by community members, will enable more comprehensive screening and support better protection and health outcomes for displaced communities.

Challenge: Despite UNHCR’s recent efforts to mainstream data-driven analysis and decision-making, many data-driven projects are still siloed within specialist units, rather than being operationalised to inform field-based work. UNHCR technical units in the field also continue to collect large amounts of data that may never be analysed beyond a very basic level, if at all.

Solution: Using data science and machine learning techniques to pilot deeper analysis of data collected by humanitarian organization HIAS on the progress of displaced households included in a specific livelihoods intervention. Using machine learning, the team will sort cases according to performance, identifying trends and common characteristics.

Expected results: Data-driven prediction of which cases are most likely to need additional support from the start of livelihoods interventions will enable a more effective response and amplify successes within livelihood interventions.

Challenge: UNHCR El Salvador collects semi-structured qualitative data from internally displaced people and community members at risk, which is analysed manually by the operation and partners. A more systematic analytical approach could improve the identification of overarching trends and patterns, which could strengthen existing advocacy efforts and community-based protection interventions.

Solution: Develop an analytical framework to support qualitative text data analysis that would provide additional insights. The project also involves community-based participatory aspects, with opportunities for at-risk youth in El Salvador to learn important data science skills.

Expected results: Better insight into overarching protection trends and drivers of protection risks, their effects on communities, and possible community-based solutions. This will facilitate advocacy efforts, support programmatic planning, and help shape effective solutions

Challenge: Cash assistance is an increasingly important part of UNHCR Peru’s support for people displaced from Venezuela. With growing needs and shrinking resources, it is critical for UNHCR to be able to provide timely, targeted, effective cash-based interventions for the most vulnerable.  

Solution: UNHCR Peru will used mixed methodology to improve the targeting of its cash-based interventions. Data analysis will identify the vulnerability profiles of families both eligible and ineligible for this assistance. Through workshops, interviews, and focus groups with partners, UNHCR, and displaced communities, the profiles that should be prioritized will be discussed. Inclusion and exclusion errors will be studied in depth.  

Expected results: More effective cash-based interventions to respond to different operational contexts, needs, and individual vulnerabilities, supporting greater wellbeing for communities most in need. 

Challenge: Displaced Venezuelans face challenges accessing information to inform their choices on arrival in Brazil. Social media can be an effective outreach tool, but UNHCR lacks evidence of the effectiveness of its social media use to both communicate critical information to Venezuelan refugees and asylum-seekers in Brazil and support protection work.

Solution: To optimize the choice architecture (how different options are presented, as well as the available tools to help people make decisions) accessible to displaced Venezuelans, UNHCR will use behavioural analytics to quantify and inform interventions – such as information campaigns – and correlate these initiatives to refugee decisions and behaviour.

Expected results: Improved, evidence-based communication will enable refugees and other conflict-affected communities in Brazil to make well-informed decisions, maximising their protection, health, education access, social and economic inclusion, and dignity.

Challenge: Jordan, a climate hotspot, hosts over 750,000 refugees. Every two years, UNHCR conducts a study to analyse and document different types of vulnerabilities across multiple sectors, from a representative sample of registered refugees. But this vulnerability assessment framework (VAF) does not include any indicator measuring climate change impacts on refugees.  

Solution: Integrate indicators into the UNHCR VAF on how climate change affects refugees in Jordan, to gather quality data that would inform the programming of UNHCR and its partners over the years to come.  

Expected results: Improved understanding of climate change impacts on the displaced in Jordan, driving better climate change related protection interventions and enhanced resilience of refugees to climate change related risks.