Estimating Internal Displacement in Ukraine Using Satellite Car Counts
Author Information
Author(s): Rufener Marie-Christine, Ofli Ferda, Fatehkia Masoomali, Weber Ingmar
Primary Institution: Qatar Computing Research Institute, Hamad Bin Khalifa University, Doha, Qatar
Hypothesis
Can internal displacement patterns be estimated from changes in car counts using satellite imagery?
Conclusion
The study found a clear East-to-West movement of cars in Ukraine following the onset of the war, indicating significant internal displacement.
Supporting Evidence
- The study collected over 1000 satellite images from 61 cities in Ukraine.
- A clear positive trend was observed between the number of cars and the population in most cities.
- The methodology allows for resource-saving and innovative approaches to estimate internal displacement.
Takeaway
The researchers used satellite images to count cars in Ukraine to figure out how many people had to leave their homes because of the war.
Methodology
The study used a computer vision model to detect and count cars from satellite images and linked these counts to population data to estimate displacements.
Potential Biases
Potential overfitting in the model could lead to inaccurate population estimates.
Limitations
Data sparsity and the reliance on satellite imagery may mask detailed dynamics of internal displacement.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website