How Institutions Can Prepare Their IT Environments for a Data Science Program
Colleges and universities use all sorts of technology — some of which is "shadow IT" not authorized by the IT department. It's easy to understand how this happens. Faculty members receive grants to conduct research, they require specialized tools, and they don't want to wait for approval from IT. Instead, they build out their own technology portfolio — without necessarily knowing whether it's safe or how best to secure it.
One solution is for the IT department to become as responsive as possible to faculty and staff needs. Another is to invest in an IT architecture and application stack built around open source software. Open source code is developed in a decentralized and collaborative way, relying on community production and peer review.
Commercial IT solutions based on open source software can be more secure than proprietary products, because they benefit from transparency and diverse input. A vibrant open source community can foster best practices in cybersecurity. It can also quickly identify and remediate security issues. And if the solution provider is an established member of the open source community, it can contribute to a more secure supply chain, tracking the code's provenance and confirming it has been thoroughly tested.
Boston University deployed a commercial solution built on an open source machine learning (ML) platform for its computer science program. The solution enables researchers to rapidly train and manage ML models either on premise or in the public cloud. It simultaneously provides an environment for an open source textbook, interactive lectures, and demonstrations. Students use a web browser to access a personalized virtual space for completing assignments and exploring ML models.
Investing in such IT strategies and capabilities can empower institutions to offer a robust data science program and establish their own data science practice. What's more, these two aspects of data science — program and practice — can enhance each other.
An effective data science program can help schools attract faculty and students alike. Likewise, the data science research — and freshly minted data scientists — that emerge from such programs can advance their data science practice, helping them gain new insights to educate students more effectively and better compete in the marketplace.
About the Author
Damien Eversmann is chief architect for education at Red Hat. Having spent the bulk of his career working in or with the public sector, he is somewhat of an expert when it comes to IT in government and higher education. Throughout his working life, he has served as a developer, system administrator, development manager, enterprise architect, and technology director.