Modeling Human Migration in Afar, Ethiopia

Gravity models

From the Garcia et. al paper, we learned the importance and benefits of utilizing gravity models when analyzing human migration. The gravity model is used to use population size and distance as push-pull factors. It was really interesting to see the importance of location as one of the major factors in migration. To begin project 3, we began by creating a gravity model for London - following a guide by Dr Ds. Essentially, we did this by calculating a distance matrix between boroughs in London, and using transport data for migration flow estimates. Utilizing the gravity model allowed us to look at flow estimates, and original flows.

Origin-Destination (OD) Matrix

This is a snippet to reveal what my OD matrix looked like. I had 11 rows and 11 columns, as I had 11 administrative subdivisions within Afar. Each number is representative of one of these administrative subdivisions. I downloaded data on internal migration flows in ethiopia, from the worldpop.data website. Next, I found the origin flows sum and destination flows sum for each adminsitrative 1 district in Ethiopia. Using these sums, I was able to create the OD matrix to have a better visualization of people migrarting between the districts. The N/A is inter-migration within districts, which was removed (migration that began in the district and ended within the same district). Each cell represents the flow centerpoints. For example, the number in the cell is the predicted migration from the centerpoint of the “row” administrative district to the centerpoint of the “column” administrative district.

Animated Migration

I used the OD matrix to create an animation of migration within Ethiopia. Looking back, it’s so croweded, I wish I produced a similar animation without the migration lines for clarity! The gravity model could be incorporated in this animation, which would be interesting as it would provide more information about the autonomous agents moving in the ABM, and additionally, to model/compare estimate flows and original flows.

Tesselation of Voronoi Polygons

Below is the voronoi polygons produced for the municipalities of Afar Zone 4, within Afar, within Ethiopia, Africa. Similar to how we created an origin/destination matrix in the London gravity model exercise, if I wanted to produce an OD matrix at these higher resolution entities, I would narrow down the data from the worldpop.org that was used to make the ABM of migration flow in Ethiopia. The voronoi projection defines the centerpoints, and to better model migration flow, I would probably modify the number of points departing from each origin. I think it would be a really good idea to incorporate transportation vehicles, as thats a better representation of what’s going on in reality.

Urbanized areas within Afar

Here, I imported the defacto settelements/ center points I located in project 1 and combined it with the synthetic population produced in project 2. This plot shows the overlap of the data.