Challenges running Data Science courses
Many institutions around the globe are now offering Data Science degrees. This is driven by the demand from Industry looking to recruit people who have an in depth understanding of the power and potential of data, who can frame complex data-driven solutions to business problems with the right analytical techniques.
However, running Data Science courses is not easy as you may think. There are many components that can make up the data science workflow. It will require a rich infrastructure and tool set that tends not to be readily accessible to universities.
We’ve summarised the 5 major challenges that Universities will face when running these courses:
- Computational power: Providing students with the right amount of compute they need for their data science and machine learning courses is challenging due to patchy infrastructure and, in many cases, lack of cloud infrastructure.
- Sharing materials: Tutors struggle to share documents, scripts, requirements files, etc. with students as they are all various levels of hardware. This results in the challenge that students from varying backgrounds might not have equal opportunities to peer students to perform their assignments. For course directors this is also a challenge: uploading and downloading exercises/assignments via existing university portals and tools (e.g. Moodle) is very time consuming and limits the amount of data they can make available for assignments.
- Data Availability: Data is held on local machines which is prone to theft and infection. This makes it very challenging to manage permissions and students’ logins requirements.
- End-to-end workflow: Students don’t get exposure to end-to-end workflows and industry standard practices such as: building APIs endpoint, tracking metrics and parameters, parallelise jobs, version modelling, etc.
- IT burdens: Free use platforms that are commonly in use by students, such as Google Colab and Azure Studio can create additional challenges when you are trying to deliver collaborative lectures or delivering to volumes of students. Infrastructure providers do not offer any technical support if something goes wrong. This means more time wasted on debugging IT issues.
How Faculty Platform helps
Faculty Platform provides a common, inclusive environment to develop Data Science skills – a managed platform to support the data science workflow. It comes with pre-installed data science toolset and best-inclass open source tools used in industry. Faculty Platform has had a remarkable impact on both lecturers and students:
They can now make the materials accessible and easy to review without wasting time downloading files and battling IT for compute resources – an added benefit if courses are run remotely. Additionally, they can now provide real life industrial scale tools to go beyond the textbook and truly prepare the students for their career in Data Science. Finally, they can rely on Faculty dedicated Engineering support for tutorials, data science advice and engineering issues.
They log in on one platform and gain immediate access to the materials and the compute power they need to help them develop the necessary skills. Students get exposure to the best tools in the industry for doing data science at scale – it allows for unlimited experimentation and learning.
If you are interested in finding out more about the Faculty Platform, contact us today.