Remote Sensing
The embedded R Markdown file shows a post-classification spatial filtering application for one land cover maps that I created for Mapping Drivers of Ecosystem Change in Colombia, a remote sensing research project I provide support to at Temple University.
The end result is an updated classification map that has a higher overall accuracy score, as small patches of potentially misclassified oil palm were detected and reclassified to natural dense forest. The decision to apply spatial filtering in this method was based on literature and observations of the classified map's accuracy. This analysis was completed in RStudio and was developed by Jerónimo Rodriguez, a PhD student at Temple University, and refined/compiled into an R Markdown script by me.