By Giacomo Falchetta and Ahmed T. Hammad
Manuscript under peer review, preprint: https://doi.org/10.21203/rs.3.rs-3916891/v1
To replicate the analysis:
- Download input data from Zenodo: https://doi.org/10.5281/zenodo.13886667
- Optional data extraction steps (processed output data are already available in the Zenodo repository):
- Adjust your working directory
- Run [lines 4-11] of workflow/sourcer.R
- Run the Javascript scripts written by the string_generator_training.R and string_generator_prediction.R files in Google Earth Engine (https://code.earthengine.google.com) and complete the export to Drive tasks to generate the output .csv files
- Run workflow/sourcer.R [lines 15-46] to train the ML model and make predictions (including figures and tables replication)
The analysis was implemented in the following environment:
- RStudio 2023.03.0
- R version 4.2.3 (2023-03-15 ucrt)
- Platform: x86_64-w64-mingw32/x64 (64-bit)
- Running under: Windows 10 x64 (build 22631)