Developing Opportunities to Automate Updates on Population Modeling Efforts for Refugee Camps

Migrant and refugee crises are highly disordered and profoundly international issues. They involve an entanglement of nations, individuals, and agencies from across the globe, each with their own concerns, objectives, and strategies. Due to the various interests of humanitarian organizations which gather data, obtaining specific data on refugee settlements and the people who live within them and is accurate, up-to-date, and consistent is difficult and often time consuming. There are many organizations which gather and publish this type of data, for example, UNHCR (United Nations High Commissioner for Refugees) provides critical emergency assistance directly to refugees and conducts research on healthcare and facilities whereas REACH Resource Center is most interested in facilitating the development of information and products designed for a variety of humanitarian efforts. ORNL is interested in opportunities to automate updates to refugee or IDP camps global data collection in support of population modeling projects; LandScan Global, LandScan HD and Population Density Tables (PDT). ORNL's population modeling platforms require the latest available statistics about camp populations, their locations, demographics, open/close status, and more. Since the data is used across multiple projects, ORNL has developed a database of camps, but there is opportunity to leverage existing in-house capabilities of PlanetSense for purposes of change detection for rapidly identifying new or abandoned camps and growth or decline of existing camps. The PlanetSense platform currently harvests, on a worldwide scale, open source VGI for Points of Interest (POI) and social media such as Twitter. This discussion will focus on opportunities to automate data collections through known open source sites such UNHCR and REACH, use the PlanetSense POI to enhance and validate the existing camp locations and metadata, and social media as indicators of change.