(DRAFT ONLY)
Self-guided work through weeks 3 & 4 of the training.digitalearthafrica.org course, followed by a use case looking at a SDG.
Self-guided teams will detect if and when land-clearing has happened within a given area, an generate mosaics of before and after the clearing event.
The datacube-ows project powers the ODC web maps at DE Africa and DE Australia.
This workshop will explain how the styling of the works, and how you can create your own style to demonstrate your own visualisation.
There will be two workshop sessions:
- Session 1: Wed 23 09:30 AEST / Tue 22 11:30 UTC
- Session 2: Wed 23 19:00 AEST / Wed 23 09:00 UTC
Prerequisites:
- Sandbox account on DE Africa or DE Australia
- Familiarity with Open Data Cube
More info: Styling HOWTO
Contact: Paul Haesler | @Paul Haesler on ODC Slack
This hack will explore scalable methods for short-term time-series predictions of remote sensing vegetation indices (e.g. NDVI, MSAVI). The output of this sprint should be one or two functioning jupyter notebooks that perform accurate short-term predictions of a nominated vegetation index, along with an account of the methods that failed. This work will hopefully inform future efforts at more advanced ecological forecasting.
Contact: Chad Burton | [email protected]
- Paper: Forecasting vegetation condition for drought early warning systems in pastoral communities in Kenya
- Example forecasting in Python
- Autoregression Models for Time Series Forecasting With Python
- ScitKit-learn-Inspired Time Series models
- https://www.machinelearningplus.com/time-series/time-series-analysis-python/
The group will decide on a topic, and use training labels (possibly from Radiant Earth http://registry.mlhub.earth/) to train a model.