My name is Rodgers Iradukunda, and I am a first-year PhD student at the University of Liverpool, where I research the cumulative effects of spatial inequalities, owing to their multidimensional nature. Before my PhD, I studied for a Master’s in Geographic Data Science at the University of Liverpool. In my master’s dissertation, I used smartphone GPS data to explore internal displacement in Ukraine caused by the ongoing Russia-Ukraine conflict. My motivation for the dissertation was to potentially inform targeted humanitarian assistance to internally displaced persons in Ukraine and other crisis-laden areas.
Summary
I set out to leverage GPS location data from smartphone applications to estimate the scale, evolution, and spatial patterns of internal population displacement in Ukraine from the start of the conflict in February 2022 to August 2022.
Using smartphone GPS data, I quantified the levels of daily population movements, identified key origin and destination areas, and estimated the extent of return movement across raions and oblasts in Ukraine. Smartphone GPS data represent a segment of the Ukrainian population. Hence, I adopted post-stratification principles to adjust the smartphone GPS-derived population estimates using WorldPop and UNHCR data.
Estimates showed that northern, eastern, and southern areas experienced the largest population losses during the early weeks of the Russian invasion. At the same time, central and western areas are recorded to have played a role as key destinations for internally displaced individuals. Inter-raion displacement stood out, indicating that official statistics may underrepresent the extent of internal displacement. The research also found that internally displaced persons tended to cluster in densely populated urban areas, underscoring the protective influence of urban infrastructure amidst conflict zones.
The key novelty of the research was to develop an approach to produce bias-adjusted estimates of population displacement from smartphone GPS data. Prior work used mobile phone data but has been mostly constrained to offer rough signals about population movements (e.g. spatial concentration), trends (e.g. increasing or decreasing) and changes (e.g. low to high).
Population Geography 