Accelerate science, research, and innovation with AWS
summary
Shows concrete tooling and results: researchers move an on‑prem Open Data Cube to AWS, use seasonal compositing of Landsat/Sentinel optical plus RADARSAT radar, generate pixel predictors, then run distributed classification (testing Amazon SageMaker and ParallelCluster) to produce a national wetland map and a planned 40‑year time series. Claims include 10x more productive ML workflows on SageMaker, use of Snowball Edge for remote collection, and elastic FSx/Lustre + EFA for HPC scaling — the pipeline scales and yields usable national mosaics.
AI-written summary · about
chapters (33)
- 0:00Intro
- 1:58Research cycle
- 2:50Science and research leadership goals
- 4:03Scientific researcher goals
- 5:00Research data needs
- 6:05Diverse needs of researchers
- 7:54Innovation for research and science on AWS
- 11:08AWS specialist teams for areas of research
- 12:51Secure Platform for Research and Science
- 16:43Amazon SageMaker Studio Notebooks
- 17:38Amazon SageMaker Canvas
- 18:32Shared or self-served clusters
- 19:49HPC with AWS ParallelCluster
- 20:41User interface
- 21:07Compute and storage can be adjusted dynamically
- 21:38Amazon FSX for Lustre
- 22:26Quantum computing on AWS
- 24:16Open Data on AWS
- 25:02Data collaboration on AWS
- 25:33Research data approach
- 28:07Cost-optimized storage
- 29:33Commitment to sustainability
- 30:46No-cost research learning pathway
- 31:28Geomatics research: Spatial tools supporting wildlife and landscape conservation Environment and Climate Change Canada (ECCC)
- 33:10What is landscape science and why is it important?
- 36:01New Paradigm Open data archives - Benefits of Using Temporal Data
- 37:31Acquiring and transforming data into needed information
- 39:43Open Data Cube Distributed processing implementation
- 40:04Seasonal compositing
- 42:52Seasonal Trend Analysis - Radar Satellites (Canadian)
- 45:03Wetland type classification - Nova Scotia, New Brunswick
- 45:54AWS Implementation to date - ECCC Geomatics
- 47:28What's next for the Geomatics Research team?
related talks
22mRunning Geospatial Workloads on AWS
28mAdvancing the future of space in the cloud
14mBattle-tested tactics to accelerate data science and ML R&D with AWS
21mAWS On Air ft. Amazon SageMaker Canvas
28mTurn Data Into Insights
40mAccelerate autonomous vehicle development
26mGeospatial Machine Learning at Scale
27mVisualizing Cloud Systems