Embracing open source dataDavid Wade, Equinor. Photo: Terje Solbakk

Embracing open source data

Equinor sees the benefits of collaboration and data-sharing. Since 2015, the company policy has been open-source for all developed software.

– Our new open-source data policy is reflected in the number of public projects hosted through GitHub, said David Wade, Principal Developer Data Science in Equinor at the DigEx 2020 conference in Oslo on January 28th.

GitHub is the world’s largest community of developers to discover, share, and build better software, from open source to private team projects.

Equinor’s public projects are hosted at their GitHub page through so-called public repositories, which have been growing exponentially in the last years, reaching 272 in Q1 2020.

Illustration: Number of Equinor projects in GitHub since 2015. Illustration: Equinor

According to Wade, Equinor sees many areas in which they, as well as other companies, have more to gain by collaboration than through competition. Since 2015, the official Equinor policy has been that all developed software by default will be open source.

Sharing data, workflows, and ideas is not necessarily done straight forward.

– Getting something for nothing sounds appealing, but there are pitfalls to be aware of. Concerns may include security issues, hidden costs, legal issues, and ownership towards project direction.

Tackling these issues comes down to professionalism, best practices, and competence among software developers, as well as open communication between them and geoscientists, according to Wade.

Examples of Equinor projects done in collaboration with other companies and that represent open source programming are:

  • OPM: Open Porous Media, reservoir simulator ++
  • segyio: Fast Python library for SEGY files
  • dlisio: Working with well logs in DLIS format
  • PyLOPS: A Linear-Operator Library for Python
  • XTGeo: Python class library for subsurface data structures
  • seismic-zfp: Compress and decompress seismic data