Data transformation
As we assembled our dataset, we noticed a few caveats in data quality worth mentioning. Very few natural disasters hit the area during our time period of interest, and these were spread across a wide geography. It’s unlikely that this factor will impact Ebola spread, but we will keep this variable in our dataset.
The source for violent incidents appears to be missing information, as known attacks on aid workers and Ebola treatment sites in Katwa are not included in the data. This suggests an incomplete data source that may not capture violent incidents at the level needed for a real-world analysis of factors influencing Ebola spread. However, for purposes of demonstrating this method, it captures enough to be potentially interesting in the analysis. In projects such as this, data quality can be questionable, as good sources are hard to find in much of the developing world.
The search for transportation routes was also questionable, yielding some insight...
