Summary: In recent years, there has been a drive toward more open, cross-disciplinary science taking center stage. This has presented a number of challenges, including providing research platforms for collaborating scientists to explore big data, develop methods, and disseminate their results to stakeholders and decision makers. We present our vision of a “data science lab” as a collaborative space where scientists (from different disciplines), stakeholders, and policy makers can create data-driven solutions to environmental science's grand challenges. We set out a clear and defined research roadmap to serve as a focal point for an international research community progressing toward a more data-driven and transparent approach to environmental data science, centered on data science labs. This includes ongoing case studies of good practice, with the infrastructural and methodological developments required to enable data science labs to support significant increase in our cross- and trans-disciplinary science capabilities. The Bigger Picture: As we move toward the need for open and cross-disciplinary science, and with an ever-increasing volume of data, there is a critical need to provide research platforms that support the wide variety of users that need to gain knowledge from this data. We present our concept of a “data science lab” as a key contribution in this area. A data science lab is a collaborative, dynamic, and tailorable platform that caters for users at varying levels of abstraction. We illustrate the concept with an initial implementation of a data science lab, drawing on our experiences in the cross-disciplinary field of environmental data science, thus aiming to develop a more data-driven and transparent approach to science. We set out a research roadmap to serve as a focal point for the international research community to take the concept forward and enable data science labs to support the ever-increasing requirement for cross- and trans-disciplinary science capabilities.
