Vegmachine API

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