The Johns Hopkins University Open Source Programs Office (OSPO), part of the Sheridan Libraries, is delighted to announce the inaugural recipients of funding through its Free and Open Source Software Project Fund (FOSSProF). This initiative, generously supported by the Alfred P. Sloan Foundation, empowers JHU faculty, researchers, students, and staff to share and collaborate on their innovative work through engagement with the open-source community.
Following a competitive review process, nine projects across five academic divisions have been selected to receive over $250,000 in FOSSProF funding. These projects embody the OSPO’s mission to promote the use of open-source software for research, education, and technology transfer for the JHU community. The awarded grants provide financial support and access to resources and mentorship, fostering the development and implementation of these innovative solutions. A summative event sharing the outcomes of the projects will be held in the Fall of 2024.
Elisabeth M. Long, Sheridan Dean of University Libraries, Archives, and Museums, says, “I’m pleased to see the inaugural FOSSProF grantees representing such a diverse range of disciplines across our university. This initiative is a wonderful showcase of the role the Open Source Programs Office plays in accelerating the development of sustainable open source software as an integral part of the research process. Software development is no longer confined to just the traditionally computational disciplines, and each of these projects is taking the next step to ensure that their work is openly available to a broad audience.”
↠ Advancing biosensor data analysis software SACMES to browser-supported format
Principal Investigator: Netz Arroyo, Associate Professor, Pharmacology and Molecular Sciences
School of Medicine
SACMES is an open source application used to perform analysis on the data outputs of continuous molecular monitors, implantable sensors that can measure and track the concentration of medically relevant molecules in the body. Funding will be used to update Python libraries, modularize the code, develop a browser-based version of SACMES, incorporate cloud data storage capabilities, and expand the capabilities of the software via AI tools to improve data visualization.
↠ Development and education of the use of R Language for implementing statistical methods in clinical research and beyond
Principal Investigator: Ruizhe Chen, Instructor, Division of Quantitative Sciences
School of Medicine
There is currently a lack of existing open-source software programs for simulating multivariate correlated zero-inflated generalized Poisson distributed count data, which can be used for modeling counts of component failures in reliability engineering, counts of adverse events in clinical trials, symptom counts of alcohol use disorder, etc. The Division of Quantitative Sciences has established a GitHub page to share code and articles for promoting quantitative methods and open-source programming in solving clinical research problems. This project aims to contribute to and promote development and education of open-source statistical programs written in the R language with a focus on implementing statistical methods.
↠ VisualScore
Principal Investigator: Max Eidinoff, Graduate Student
Peabody Institute
VisualScore is a planned web app containing only the most essential and unique notation features for composers and music publishers. This web app will combine fast music notation functionality with the unlimited possibilities of graphic design tools, allowing composers to create truly anything they imagine within a single program. The project has also received a Peabody Launch Grant, and is currently working on UI/UX design and development of an initial prototype of the software.
↠ Poseidon-Viewer enhancement and documentation
Principal Investigator: Thomas Haine, Professor, Earth and Planetary Sciences
Krieger School of Arts & Sciences
This project will improve software quality, increase functionality, document, and disseminate the Poseidon-Viewer application. This tool provides easy access to massive ocean circulation model simulations, as part of the Poseidon Project, which runs on the JHU SciServer resource as part of IDIES. The Poseidon-visualization tool provides very easy access to the massive Poseidon Project ocean general circulation model solutions. The tool aligns with the vision to “democratize the ocean circulation model data” in the Poseidon Project; namely, to make the data available with minimal barriers to access to anyone with a web browser.
↠ Cloud migration and scaling for community use of Rodent Automated and Integrated Learning (RAIL) software system
Project Lead: Barbara Holt, Research Technician, Mysore Lab
Principal Investigator: Shreesh Mysore, Associate Professor, Psychological and Brain Sciences
Krieger School of Arts and Sciences
Neuroscience research has seen a recent explosion in interest in visually guided behavior in rodents. There is a growing need for a high-throughput, touchscreen-based system – hardware “boxes” and control software – for parallelized training of large cohorts of freely moving rodents for both basic research and cognitive testing for drug development. Existing commercial systems are prohibitively expensive, have limited adaptability (proprietary), and are difficult to scale-up. Mysore Lab has developed the scalable, open-source RAIL hardware system using IoT parts at a fraction of the commercial cost. There is still a critical need for software to control parallelized training in banks of these behavior boxes, with access to source code for flexible use. We have developed the open-source RAIL software to address this community need; however, it currently exists in prototype.. This project will move RAIL to a user-friendly production environment available to the community.
↠ Expanding and refining Steamroller, a flexible Python library for scalable, reproducible empirical experiments
Principal Investigator: Thomas Lippincott, Associate Research Professor, Humanities Institute and Computer Science
Whiting School of Engineering
Over the past several years, faculty and students have made significant use of Steamroller, a Python framework designed to handle many of the challenges to effective design and execution of computational experiments. Due to the history of the project, and the recent evolution of grid environments at JHU, Steamroller is currently limited to a narrow subset of features supported by CS-affiliated grids, and has accumulated a list of planned augmentations. This proposal aims to resolve both of these issues,and will allow the project team to devote a significant amount of effort over the coming semester to ensuring Steamroller provides a stable, productive backbone for computational research at Hopkins.
↠ Reproducibility standards for Economics
Principal Investigator: Alan Lujan, Visiting Assistant Research Professor, Economics
Co-Principal Investigator: Chris Carroll, Professor, Economics
Krieger School of Arts and Sciences
The Economics profession has fallen behind other fields in software development and reproducibility practices. To address this problem, Econ-ARK has been working on REMARK, a set of standards and tools for reproducibility in our own work in Economics. The objective of REMARKs is to be self-contained and complete projects, whose contents should be executable by anyone on any computer that meets a minimal set of requirements and software. The goal of this project is to expand the REMARK standard to allow for the use of modern scientific publishing technologies such as Jupyter notebooks and MyST markdown, and extend the REMARK standard to integrate with the Open Journals infrastructure.
↠ Driving precision medicine forward: large-scale transfer learning for multi-institutional health data via extension of the BayesBridge package
Principal Investigator: Akihiko Nishimura, Assistant Professor, Biostatistics
Bloomberg School of Public Health
This project aims to deliver a high-performance implementation of Bayesian skew-shrinkage regression models and integrate it into the HADES software suite developed by the Observational Health Data Sciences and Informatics (OHDSI) consortium. The proposed statistical tool will address one of OHDSI’s most urgent needs: to transfer information obtained from training a model on a larger database to a smaller database, without sharing patient level data.
↠ Implementing and integrating uncertainty-aware machine learning methods in UQpy
Principal Investigator: Michael Shields, Associate Professor, Civil and Systems Engineering
Whiting School of Engineering
Currently, several open-source packages exist that address the topics of either machine learning (ML) or uncertainty quantification (UQ). Only recently, new scientific advances have resulted in methodologies that lie at the intersection of the two fields. The focus of this project will be to survey, gather and implement methods that lie at the intersection of ML and UQ and create a new module in the UQpy software. This new module will provide a unified toolbox that can seamlessly integrate UQ tools existing in UQpy with ML capabilities that exist across a set of disparate Python tools.