An overview of resources for the Machine Ocean project.
Machine Learning Libraries
The project will rely primarily on the following open source machine learning libraries:
Other libraries may be added as needed, but the project will avoid proprietary software.
Source Code
All code used for Machine Ocean is tracked in various git repositories.
To make contributions, please fork the source code and submit pull requests via GitHub.
Machine Ocean source code should be released under the Apache 2.0 License as a rule,
with the MIT License as a secondary option if strictly necessary.
This matches the licensing policy at MET Norway.
Data Sources
An overview of the data sources that is intended for use with the Machine Ocean project.
Measurements from Östergarnsholm Island
| Storage |
This data is hosted by Uppsala University |
| Type |
Measurements |
| Access |
The datasets are archived and open access for non-commercial use |
| URL |
www.icos-sweden.se/station_ostergarnsholm.html |
| Format |
ASCII plain text, MATLAB |
| Size |
50–100 gigabytes |
Measurements from Swedish Lakes
ASIT data from Martha’s Vineyard Coastal Observatory
| Storage |
This data is hosted by WHOI |
| Type |
Measurements |
| Access |
The datasets are archived and open access for non-commercial use |
| URL |
www.whoi.edu/website/mvco/home |
| Format |
ASCII plain text |
| Size |
50–100 gigabytes |
Data from RASEX, MBL, CBLAST, CLIMODE, SPURS-1 and SPURS-2 Campaigns
| Storage |
This data is hosted by WHOI |
| Type |
Measurements |
| Access |
The datasets will be made publicly available |
| URL |
www2.whoi.edu/site/casimas |
| Format |
ASCII plain text, MATLAB |
| Size |
50–100 gigabytes |
Data from Ocean Observatories Initiative (OOI)
| Storage |
This data is hosted by the Ocean Observatories Initiative |
| Type |
Measurements |
| Access |
The datasets are archived and open access |
| URL |
oceanobservatories.org |
| Format |
NetCDF, CSV, JSON |
| Size |
Approximately 200 terabyte (raw data archive) |
Data from Copernicus Sentinel I, II, and III