We seek to develop and utilize innovative, interdisciplinary tools for data analysis and visualization, to support the scholarly pursuits of ourselves and others and to build a strong community engaged in participatory inquiry.
The DXL was formed in the Spring of 2015 as a result of funding from the Gordon and Betty Moore Foundation through the Data Driven Discovery Investigator program in a grant to Matthew Turk, under the title “Changing How We Conduct Inquiry” (proposal and presentation) We are also grateful for funding from the National Science Foundation under the SI2-SSE program to develop the yt project (proposal).
In our vision of the future of data analysis, scientists should no longer have to struggle to understand their data. They should be struggling to understand their discoveries. Moore’s Law will push processing power another order of magnitude in the next five years; we aim to make discovery an order of magnitude faster.
The difficulty in simply asking questions of data has begun to dominate the cost of conducting scientific inquiry. In our lab, we focus on reducing the technical friction between a scientist and data, on enabling new classes of questions to be asked, and on empowering scientists to explore data sets without apprehension of scale. This will take the form not only of technical development, but engaging in discussions of community development, diversity growth, and the sustainability of scientific software. More than developing a software package or set of packages, this is an entire approach to scientific inquiry.
Crucial to developing and implementing this approach to scientific inquiry is that our products must be publicly developed, openly and freely available, and supported by community engagement that is designed to empower and support individuals. This extends both within the lab and beyond it; within the lab, we will foster a collaborative lab environment in which individuals feel comfortable sharing data, struggles, successes, and technologies.
The three key goals we intend to accomplish:
- Broader: We believe that scientific analysis should be interdisciplinary. Technologies should be spread between domains.
- Deeper: Individuals should be able to ask questions of their data that are complex, detailed, and non-trivial with ease.
- Further: The process of applying analysis to data should be simplified and made accessible to all communities, through the utilization of data sharing and analysis systems.
Finally, a crucial aspect of this is the success of individual lab members pursuing their own scholarly inquiry. The advancement of tools and technologies will be coupled with their application; postdocs and graduate students will be encouraged to continue projects that generate data, for instance in studying the formation of stars and galaxies, seismic wave propagation, structural integrity of manufactured materials, digital humanities, and other scholarly pursuits.
Guiding this mission are a number of values that are core to our operations.
- Diversity: A diversity of many voices is essential for not only the success of the scholarly community but the well-being of any individual collaboration.
- Collaboration: Community and collaboration are key to sustaining success in scholarly inquiry. We will collaborate both within the lab and with external individuals.
- Empowerment & development: Software can be a tool to provide self-direction for researchers, enabling self-development and growth.
- Collegiality & humanity: A healthy community is one in which respect, kindness, and empathy are guiding forces, as well as appropriate humor, collegiality and friendship.
- Technology: Technology can be applied to inquiry, but is not necessarily an end in and of itself.
- Application: Development of tools and infrastructure should be guided by those who apply it, and the best way to ensure we are guided is to be leading the charge in conducting scholarly inquiry for our own research goals.
- Sharing: Data, code and other digital research objects should be shared whenever possible, in ways that enable reuse and discovery.
DXL and yt¶
The Data Exploration Lab’s goals are independent of those of the yt project. However,
yt may be an appropriate platform to
guide them or develop them, particularly as there is a great deal of internal
yt in the DXL. As such,
yt can be explored as a
delivery platform for algorithms, interdisciplinary collaborations, and
advanced inquiry methods before other platforms.
Whenever development or application of
yt is conducted within the lab, it
will follow the community standards of the
yt community, including code
review, YTEP preparation, and participation in the mainline code repository.