A Jupiter notebook is a research document containing live code, annotated write-ups and descriptions of analyses or research.
A Jupyter notebook aims to make research re-producible and re-usable by offering a common programmatic interface whereby all diagrams, tables, plots are results of inline executed code.
Being able to read the code that generates results adds full transparency to research as well as promotes the adoption and re-use of that research.
This repository houses various Jupyter notebooks from the open data cube community. A varied community of remote sensing, GIS, and earth observation specialists/researchers.
The open data cube is used in these notebooks to query large data-sets for time series rasters on which analysis is conducted.
-
Loading Data
This notebook details retrieval of data from the
open data cube
. Topics include establishing a connection to the data cube, defining what data gets loaded, and a high level description of thexarray
object returned by the load operationLink: datacube load tutorial
Data:GPM
-
Forest Degradation using Linear Regression Analysis
This notebook runs regression on an NDVI time series. Slope of a regressed line is used as proxy to determine vegetation gain or loss. Based on the publication Assessment of Forest Degradation in Vietnam Using Landsat Time Series Data by Vogelmann Et al.
Link: forest degredation
Data:Landsat 7 Collection 1
-
Cloud Analysis on Landsat
This notebook compiles a series of summaries and visualizations regarding cloud coverage on landsat imagery.
Link: cloud analysis
Data:Landsat 7 Collection 1
-
Coastal Change This directory houses notebooks on coastal analysis, including a very simple coastline classifier.
Link: coastal_change
Data:Landsat 7 Collection 1
-
Land Change Detection on ALOS imagery This directory houses a change detection case-study in vietnam on ALOS imagery.
Link: land change
Data:ALOS2 PALSAR2 SCANSAR