NCSI: Processing Multispectral Satellite Imagery for Marine and Coastal Biodiversity Assessment

Status	Completed
Mentor Name	Tylar Murray
Mentor's XSEDE Affiliation	Campus Champion, Startup Allocation
Mentor Has Been in XSEDE Community	1-2 years
Project Title	Processing Multispectral Satellite Imagery for Marine and Coastal Biodiversity Assessment
Summary	Develop optimized processing of satellite imagery for oceanographic and coastal research. Collaborate with research staff to analyze and publish taxonomic occurrence, in situ oceanographic sensor, and gridded satellite imagery data at large scale. Support existing community efforts through code reviews, expanding test coverage, issue reporting, and documentation.
Job Description	The student will help develop and deploy processing of multispectral satellite imagery for the purpose of marine biodiversity assessment. Some algorithms involve optical calculations based physical parameters such as solar angles and atmospheric pressure fetched from external sources or theoretically estimated. Others apply machine learning classifiers trained on ground truth data to determine the benthic or coastal habitat class. Processing is coded primarily in python and jobs are orchestrated within an apache-airflow cluster running atop XSEDE's jetstream and USF compute resources. The student will aid with testing and improved integration of these existing systems with XSEDE's resources by leveraging singularity, docker, or other technologies.
Computational Resources	Processing will be executed within jetstream instances under our existing startup allocation. Forthcoming exploration of additional resources and requests for additional resources under research and/or education allocations will allow for expanded capabilities within jetstream or other XSEDE compute resources. The student will be involved in discussions and works with XSEDE ECSS provided under our allocations.
Contribution to Community
Position Type	Learner
Training Plan	Following the guided apprenticeship model, the student will be expected to choose specific directions of interest within the scope of the project. Student tasks will be centered around contribution to open source software and documentation. The student will be exposed to git workflows, issue reporting via JIRA and/or github, relevant discussion forums, and direct interaction with researchers. On top of this self-directed foundation, periodic meetings to assess task progress will supplement ad hoc discussions, meetings, and asynchronous communications. These meetings will be used to tailor training tasks for the student which may incorporate online course material, code review, and practical application of skills learned.
Student Prerequisites/Conditions/Qualifications	A familiarity or interest in learning bash, python, github, *nix
Duration	Semester
Start Date	01/06/2020
End Date	03/27/2020