An eScience Deep Dive is a short-term project done jointly between one or more eScience data scientists or research scientists and other researchers within or outside the University of Washington. Deep Dives are projects eScience staff consider to be of special importance because they relate to a strategic partnership, are meant to be the foundation for a joint grant proposal, or are otherwise-motivated short-term, deep engagements.  

Because of their special role, Deep Dive projects provide communication to eScience management, staff, and the broader data science community about their progress.

A Deep Dive is initiated based on the interests and expertise of eScience research staff.

Requirements for a Deep Dive

Deep Dives may take a variety of forms and durations but are unified by common reporting expectations, a commitment to working reproducibly, and clear identification of deliverables (although actual deliverables will vary).

The requirements for a Deep Dive are:

  • Create a project web page that documents:
    • The motivation for and specifics of the problem addressed.
    • Deliverables and how they advance the solution to the problem.
    • Timing of the deliverables. We expect that this is typically one to two quarters.
    • Identify the participants in the Deep Dive.
    • Results achieved (at deliverable dates).
  • Link to the above web page from the “Projects” section of the eScience Deep Dive web page.
  • All code developed during the Deep Dive should be in a GitHub repository that should be under the eScience organization or have a fork under the eScience organization. The code should be licensed under an open source license and should be linked from the project web page.
  • Optionally, a presentation of the results at an eScience-sponsored seminar.

Deep Dive Projects

  • xState: State Analysis for Gene Expression: This deep dive aims to demonstrate the value of state analysis for understanding the behavior of biological cells by creating computational methods for state analysis of gene expression and building open source tools that support these methods.
  • PhenoSat:This deep dive expands our ability to observe ecological and physical processes in arctic and boreal lakes by leveraging the expanding catalog of high-resolution remote sensing imagery (Planet Labs) in combination with serverless computing for data processing and analysis.