Our final project examines migration between different districts within our selected countries. To start, we read “Dr Ds Idiots Guide to Spatial Interaction Modelling for Dummies” and proceeded to produce a gravity model according to commuting data from the 2001 England and Wales Census. By calculating flows between boroughs and defining a gravity model, we outlined the process we then completed in our selected areas of interest.

Our origin-destination (od) matrix is a square matrix representing the distances from each destination and origin. The centerpoints of each district are the origins and destinations modeled in our od matrix, and each path between districts is represented in a straight line using kilometers. The only connections we chose to delete were self-connections (distance equal to 0) resulting in a diagonal group of ‘NA’ values. This odm matrix is a 9x9 matrix representing the 9 administrative districts in Gabon.

The interactions between each district in Gabon are within a clean table, which means we are now able to produce analyses and visuals. The next maps of Gabon show color gradients dependent on the amount of inmigration and outmigration respectively.

A plot displaying outmigration values:

Additionally, I produced a plot showing the interconnections between districts pn the most linear path and according to the geometric center of each district’s polygon. Here is a visual tessellation of connections between counties in Gabon.

The animation we produced is clearly used for instructive purposes, but we could have implemented our OD matrix and gravity model into our simulation. This would change a lot about our current animation, which features straight lines and equivalent arrival/departure times. The gravity model would adjust our lines between origins and destinations to draw closer to population and further from remote areas. Apart from implementing these two pieces, modifying the time variable to reduce travel during night and accounting for elevation, roadways, or land use to influence paths traveled would surely increase the credibility of our migration model. Lastly, there was data in the Gabon data set on which households owned a vehicle; this information could have been used to release multiple points from each origin, some with vehicles and some without, to compare travel times when using public/personal transportation versus other modes of movement.