Who Said Wrangling Geospatial Data at Scale was Easy?
summary
Shows concrete pipelines: converting Census shapefiles into partitioned Parquet with census_parquet and Dask‑GeoPandas for scale, using STAC/stack to pull tiles, then labeling rasters with xarray-spatial and rasterizing/aggregating with data shader. Demonstrates a CUDA hillshade and viewshed that run on GPUs via RTXpy/CuPy and numba and reports a ~6000x speedup on a hotspots kernel. Everything is tied to planetary data on the Microsoft Planetary Computer and the examples run end to end.
AI-written summary · about
chapters (17)
- 0:00Introduction
- 1:33Vertical Scaling Example
- 2:30Brendan Collins Introduction
- 3:07About MakePath
- 3:36MakePath Blog
- 3:52Sophia Yang
- 4:14Natalie Odell
- 5:09Data Formats
- 5:59Vector Data
- 7:21DasGeopandas
- 10:02Raster Data
- 11:41No Labels
- 12:47Xray Spatial
- 16:13Data Shader
- 18:52Dasc
- 19:08Kupai
- 19:55Planetary Computer
related talks
25mA Journey Through a Geospatial Data Pipeline: From Raw Coordinates to Actionable Insights
40mData-centric exploration using intake, dask, hvplot, and more
28mProcessing Large Geospatial Datasets with Dask & Xarray
28mA pythonista’s pipeline for large scale geospatial analytics
31mUnpacking the geospatial engineering toolbox – an overview of data science techniques for spatial da
31mOpenDataCube - Petabyte Scale Satellite Imagery Processing in Python
35mVisualizing research data: Challenges of combining different datasources
01mA Practical Introduction to Geospatial Python