Vegmachine API
Analysis
Client
- Ecology researcher at the University of Sydney
Purpose
- To provide researcher with 30+ years worth of clean data for modeling.
Approach
Scraped 30 years of vegetation data , comparing discrepancies between scrapes & identifying correct data by overlapping with BOM records.
Used a high-performance computing cluster to download, crop and process 30 years worth of satellite imagery as input for machine learning.
Provided client with documentation, instructions for rerunning the scrapes if needed & data. ## Outcome
Client has subsequently engaged SIH for a comissioned project.
Key tools
- R: Rmarkdown, tidyverse, sf, raster, tmap, jsonlite, data.table; Git + GitHub