Renku – the ETH platform for digital reproducibility in science

The general lack of reproducibility is a well-known problem in academic research, which also extends to the digital aspects of research: all too often, independent researchers cannot even reliably reproduce the purely computational steps of data models or computer simulations.
The ease of cross-institutional collaboration and the repurposing of research artefacts (programming code, raw data, etc.) are closely linked to digital reproducibility.
This gives rise to the objectives that the Swiss Data Science Center (SDSC) aims to achieve with its RENKU platform: digital reproducibility, reusability of code and data and simplicity of collaboration.
By adhering to the following best practices, it is already possible largely to achieve these goals today:

• Code versioning
• Data versioning
• Logging and versioning of the computing environment (operating system, libraries, etc.)
• Logging of all process steps

These best practices are individually addressed in freely available software solutions and web applications – many of them related to software development. However, the simultaneous use of all relevant libraries, tools and platforms involves too much additional work for most researchers. In his presentation, Dr Andreas Bleuler shows how RENKU can be used to reduce this effort drastically by combining technologies such as Git, GitLab, Docker or Jupyterlab into one easy-to-use platform.