Users and Usage#

The demand for OSS is everywhere, though its users and usage is not always apparent. Recent reports estimate that 95-97%1,2 of mainstream IT organisations leverage nontrivial OSS assets within their mission-critical IT portfolios, whether they know it or not. Meanwhile, internal government policy is emerging in the USA and EU, “encouraging and leveraging the transformative, innovative and collaborative power of open source, its principles and development practices”3, with more-and-more government agencies adopting OSS within their operations – including many of the identified projects within this report. When it comes to sustainability, however, there is much potential yet to be realised. While tracking outcomes and impacts associated with OSS remains challenging, the following trends highlight the importance of open source within environmental sustainability for fellow contributors, end users, and greater society.

Dependencies#

Since open source is free to acquire and freely available, its ultimate use is difficult to track. Much of the open source usage arises in integrating libraries or APIs as dependencies of other software projects. Importantly, this dependency on OSS will not be apparent to many users; especially in closed-source software, where the dependency on OSS is not always made evident. Users act as evangelists by sharing OSS with other users or organisations that can benefit from it, and in return provide valuable feedback and expertise.

Project usage data from public software development and version control platforms remains scarce. GitHub, unfortunately, offers little support in compiling accurate statistics. Additionally, statistics on package manager downloads are not universally available and must be obtained through the various platforms and their APIs. While this is technically possible, it was not feasible given the study’s limited resources and timeframe. However, with the limited data obtained from the Python ecosystem, it was possible to identify individual projects with a high circulation but a low DDS score. Here projects like cfgrib, sentinelhub-py or Meteostat stand out. Those projects widely used and depend highly on the goodwill of a single developer. The median DDS of 0.436 over the 50 most used Python projects indicates that the burden still lies with a few strong contributors leading the development.

The user community of major projects in energy and battery modelling, such as PyBaMM and PyPSA, is split relatively evenly between academia and industry, with fewer users coming from NGOs and independent consultancies. In some cases, industry can drive explosive user growth. For example, over a five-year period, pvlib-python saw thousands of downloads per month. This was driven primarily by several commercial firms who integrated the library into their software products, effectively distributing pvlib-python to their clients.