High-resolution population estimates improve understanding of displacement in South Sudan

March 29, 2021

GRID3 has updated its high-resolution gridded population estimates which provides timely geospatial data for the Government of South Sudan and other decision-makers. The dataset is an current estimation of population counts for all of South Sudan, the newest country in the world, which has not collected population data in its entire territory since 2008, before the country’s independence. 

Since gaining its independence in 2011, South Sudan has been affected by instability and conflict, resulting in internal displacement that has prevented the government from obtaining a comprehensive, demographic picture of the country. Through the top-down modelling approach which used the Random Forest methodology, GRID3 South Sudan technical experts produced the estimates by combining 2008 county-level census data collected by the National Bureau of Statistics (NBS) with field-collected displacement data from IOM (through its Displacement Tracking Matrix) and UNHCR, the UN Refugee Agency. The latter used information on where displaced people have moved to and from in order to estimate displacement. 

Producing the data

The first step in producing the population estimates was to create a map that demonstrates the location of displaced people in South Sudan at high-resolution (approximately 100 x 100 metre grid cells). This was done by disaggregating IOM and UNHCR data on the locations of internally displaced persons among buildings (using Digitize Africa building footprints from Ecopia.AI powered by Maxar). The next step was to augment the map with information on where displaced people were originally located. Accurately pinpointing the place of origin can be difficult, as various communities will often use different settlement and administrative unit names. However, the GRID3 South Sudan team used the Armed Conflict Location Events Database to estimate the number of people displaced from each area of the country. In the final step, the GRID3 South Sudan team added census projections from NBS to the map.

Grid cell level population estimates accounting for displacement across Juba, South Sudan.

 The highest population counts, seen in red, show areas with large numbers of internally displaced persons (IDPs). Notably, in the southwest of Juba are the UN Protection of Civilian sites.

“Our approach ensures that the population estimates are not only based on the best national demographic numbers, but also reflect distribution across the country on a high spatial scale. Having this local-level information can be particularly useful to field teams that require accurate population numbers for development interventions, such as vaccination campaigns,”

says Dr. Claire Dooley, the lead GRID3 South Sudan modeller. 

The GRID3 South Sudan team is now working with NBS to design a population estimation survey that will lead to the creation of an even more accurate population dataset. The team will also undertake a range of data production activities in order to support government and humanitarian planning, part of a wider effort to ensure a robust geospatial data future for South Sudan.

The South Sudan population estimates can be accessed through our Data page.

More news

May 13, 2021
Physical distancing datasets can help control COVID-19 spread in sub-Saharan Africa’s urban areas

GRID3 has released 50 new datasets that can support countries looking to use nuanced, data-driven approaches to controlling the spread of the novel coronavirus. These data provide information on urban areas in sub-Saharan Africa where population density and the spatial arrangement of buildings may make it difficult to maintain physical distancing (also known as “social distancing”). 

March 18, 2021
Nigeria’s young data science talents seize new geospatial data opportunities

Data Science Nigeria (DSN) and GRID3 Nigeria share a common goal: increasing the number of geospatial data experts across the country to support Nigeria’s national development agenda. For the second year in a row, GRID3 was invited to attend DSN’s 2020 Artificial Intelligence (AI) Bootcamp for a series of workshops focused on geospatial data analysis, tools and applications.

February 16, 2021
Radio campaigns use geospatial data to reduce barriers to contraceptive use in Zambia

Using risk data provided by GRID3 in collaboration with Fraym, DMI was able to refine a Social and Behaviour Change Communication (SBCC) strategy which included running a radio campaigns for specific districts with the aim of addressing barriers to the uptake of contraceptives, such as misconceptions about their safety, low support from men, and stigmatisation of young people accessing contraceptives.