Software and Data¶

Recommended Computing Environment¶

Choose either R or Python for the main workflow, but all analyses should be reproducible from scripts, notebooks, or a documented pipeline.

R Packages¶

• sf for vector spatial data.
• terra or stars for raster data.
• dplyr, tidyr, and ggplot2 for data wrangling and plotting.
• targets or renv for reproducibility.
• SDM and biodiversity packages selected as needed for course exercises.

Python Packages¶

• geopandas for vector spatial data.
• rasterio, rioxarray, and xarray for raster data.
• pandas, numpy, and scikit-learn for data processing and modeling.
• matplotlib, seaborn, or plotnine for figures.
• conda, mamba, uv, or venv for environment management.

Candidate Data Sources¶

• GBIF occurrence records.
• iNaturalist research-grade observations.
• VertNet or other taxon-specific occurrence repositories.
• WorldClim, CHELSA, PRISM, or ERA5 climate data.
• MODIS, Landsat, Sentinel, or derived remote sensing products.
• Soil, terrain, hydrology, land cover, and protected area datasets.
• OpenTree, VertLife, Fish Tree of Life, or clade-specific phylogenies.
• TRY, BIEN, GIFT, or other trait and range databases.

Data Management Expectations¶

• Keep raw data separate from processed data.
• Record download dates, licenses, citations, and filtering decisions.
• Use relative paths inside project folders.
• Never manually edit raw data files without documenting the change.