Spatial data science in the Tidyverse
summary
Explains how sf and ggplot2::geom_sf make vector tidy spatial analysis easy but fail on large rasters, time series and out‑of‑memory data; shows stars implementing raster and vector data cubes with lazy I/O, downsampling proxy reads and NDVI-on-demand so a 10,000×10,000 Sentinel image plots in seconds instead of minutes; describes NetCDF handling and cloud/off‑disk workflows via tidync, and demonstrates gridded aggregation of hurricane tracks into equal‑area cells for density counts.
AI-written summary · about
chapters (10)
- 0:00Intro
- 1:16The good news
- 3:48What about time, and raster?
- 8:02support in base and dplyr
- 9:19Raster-and vector data cubes: stars
- 10:13Raster types
- 13:34Cloud proxies; data cube views
- 14:11None-data cube extensions
- 14:23Hurricane dataset
- 16:34Concluding
related talks
05mExploratory Spatial Data Analysis in the tidyverse
25mEverything but maps with spatial tools
01h 01m50 ways to show your data
26mWho Said Wrangling Geospatial Data at Scale was Easy?
18mFrom idea to code to image: Creative data visualizations in R
48mData Visualization in Practice
40mData-centric exploration using intake, dask, hvplot, and more
30mUnlocking Geospatial Data Science: A Strategic Overview