Data science is one of the most lucrative jobs of the current times. Year after year, the number of aspirants lining up to join this field is increasing. Most of these candidates undergo upskilling programs offered by edtech platforms or offline coaching institutes to prepare for the role. But given the growing importance of this field across sectors and the enthusiasm among people to pursue it as a full-fledged career, is it time for colleges and universities to offer data science as a separate engineering discipline?

Analytics India Magazine spoke to a few industry veterans and academicians to understand the pros and cons of taking this path.

How does a B tech in Data Science help?

Venkat Raman, co-founder, Aryama Labs, says, “I would say it is a positive development. Academia and industry alike are now starting to realise that short 3 months/6 months/1 year courses are not rigorous enough to make anyone a thorough data scientist. In-depth statistics and mathematics knowledge is essential to succeed in data science. To develop this in-depth knowledge, short time frame courses are simply not enough.”

• A longer course can lay down a solid base in fundamentals

Usually, students and aspirants who want to pursue a career as a data scientist look for short courses of 3 to 6 months from training institutes that provide such data science training programs. Though these courses do help, a full-fledged four-year undergraduate program in data science from recognised universities will create a better foundation. It will take students from basic concepts to advanced theories of data science with practical exposure without any time crunch that short training courses often lack.

IITs have been at the forefront of launching such programmes. IIT Guwahati has recently introduced a B Tech programme in Data Science and Artificial Intelligence from the academic year 2021-2022. The first batch of 20 students will be admitted based on JEE Advanced 2021 counselling.

“Having a dedicated undergraduate program helps in training students in a wholesome manner. In a well-designed undergraduate program, the foundational courses will cover the essentials of DS and AI, whereas the advanced and elective courses broaden their perspective in applying DS and AI across several domains,” Professor Ratnajit Bhattacharjee, Department of Electronics and Electrical Engineering, IIT Guwahati and Dr Ashish Anand, Associate Professor, Department of Computer Science and Engineering and Associate Faculty, Mehta Family School of Data Science and Artificial Intelligence, IIT Guwahati, jointly wrote in an email response to AIM. 

• Headstart

CBSE recently announced that for the 2021-2022 session, it would introduce coding as a topic for students in classes 6th to 8th and data science for students in 8th to 12th. If school students are already getting a taste of the flavour of data science and AI, it will make sense to offer an undergraduate degree in data science for students who want to make a career in this. It will give them a good head start over students getting introduced to it at a postgraduate level.

But is it too niche?

Data science requires knowledge of different subjects, and providing an undergraduate degree in it can be way too niche, some feel. Anubhab Chatterjee, Researcher at Wipro AI Labs, says, “I don’t believe in Data Science being offered at an undergraduate level as a specific course in itself. Data Science is a vast domain and deals with numerous subjects such as mathematics, statistics, computer science and so on and so forth. Only with a clear basic understanding of these subjects can one indulge in the methods of Data Science.” 

“Data Science is not a subject in itself; it’s the method of handling, processing and drawing meaningful insights from data. Since data is something that is present in subjects of all domains, data science can hence be a part of any such curriculum that sees the need for it. Data science should be introduced not as a subject but rather as an arsenal of techniques that help everyone tackle data in their respective domains,” he adds.

How to introduce data science in the curriculum?

What industry thinks

A B Tech in data science can be a great course to create the data scientists of tomorrow, but universities have to be cautious in how to plan such a curriculum for students who have passed out from school. Raman, who frequently hires data scientists, points out certain things the universities should do:

• Make sure students are really thorough in the statistical and mathematical concepts.
• Please teach industry-relevant topics too. Often, academia falls behind in keeping up with the industry. When the students appear for data science job interviews, they don’t have an employable skill set. So, the onus of teaching these students industry-relevant things falls squarely on the shoulders of the industry.
• Increase industry exposure and encourage students to take up more internships. Universities should allocate at least 45% of the course for internships.

What academia thinks

Professor Ratnajit Bhattacharjee and Dr Ashish Anand point out:

• While designing the curriculum, identify the objectives properly. A well-balanced curriculum providing foundational and practical knowledge in data science requires, primarily, courses in computer science, mathematics and statistics and electrical and electronics engineering. Computer science courses help develop programming, big data management and analytics skills and enhance understanding of the relevant issues in computer systems and architecture. Mathematics and Statistics courses introduce mathematical and statistical tools to work with a large amount of data.

• Full-fledged undergraduate and postgraduate program – The advantage of such a program is that trained students will have a robust theoretical understanding (from basics to advanced concepts) and practical training. This will help them to adapt to any specific sector or contribute to the core data science and AI research.
• A selected set of 5-6 courses as a minor in Data Science and AI – Students from other disciplines can take advantage of such a minor program.
• Departments introducing or adapting a few courses in their curriculum to train students to apply data science and AI tools and techniques to the specific domain problems – students will be introduced to the basic concepts and how to use them in problems of a particular domain/department.
• Develop a well-balanced course structure with appropriate emphasis on learning relevant theoretical fundamentals and practical training with hands-on sessions. On top of this, include a few advanced courses along with a few courses of interdisciplinary nature based on the strength of the university/college/department.
• Give wholesome training instead of picking courses based on hot keywords