How Academia Can Fill the Data Science Employability Gap
A latest research by the World Economic Forum on The Future of Jobs in 2018. In the next 4 years, more than 75 million jobs may be lost as companies shift to more automation. But the projections have an upside: 133 million new jobs will emerge during that period, as a new division of labour between people and machines. The Demand for data Scientists is increasing significantly and specialists in Data Science – Artificial Intelligence, Machine Learning, Computer Vision and Natural Language Processing-based technologies are expected to be in greater demand over the next few years.
The Indian context, technology industry is battling with the challenge of low employability of fresh college graduates. A couple of decades back, with little peripheral knowledge and skills, fresh graduates would be on-boarded by companies, and in due course, organizations would make them market-ready. But now, the market has become more demanding. New technologies are coming up at a neck-break speed, and a lot of businesses today offer very specific and niche solutions that demand extremely specific skills.
Today, organizations don’t necessarily have the time, patience, and resources to train fresh graduates on niche skills and make them employable. In such a scenario, the onus lies on the educational institutions to fill in the gap.
Data Analytics and data science are some of the fastest-growing fields in recent times. Skill sets such as Data Architect, Data Scientist, Data Analyst, Data Engineer, Data Visualizer etc. are in high demand. However, the market demand for these skills is still unfulfilled. According to a 2018 LinkedIn workforce report, there was a shortage of 151,717 resources with data science skills. The gap doesn’t appear to close in the times to come if deliberate steps are not taken today. Currently, in the US, only one-third of the top 100 global universities provide data science degrees. The number could be even lesser in India.
A Data Scientist needs skills that can be broadly classified into three areas. Firstly, knowledge of basic tools like R, Python, and SQL. Secondly, knowledge of statistics. Thirdly, understanding of Machine Learning, which includes knowledge about random forests, K- nearest mean, etc.
Nowadays, data science training courses are mushrooming, but the real challenge is, the gap in what industry needs and what institutes are producing – a quick fix short duration theoretical and class-room led approach. Data Scientist is not just another software programer, but requires a use-case drive problem solving and research driven approach. It needs to be supplemented by the knowledge and tools, like Rubics.io, or SAS that enable data enthusiasts to define the scientific goals, prepare data, build models, explore data, test hypnotists produce results and create an industry or subject specific solution
Let us deep dive into how colleges can bridge the skill gap that is increasing to an alarming level.
- Firstly, the call of the hour is for educational institutions to set up a Data Science Center of Excellence. This can be an overarching governing body whose ownership can be jointly managed by both the institution as well as the COE sponsor. It is recommended that the COE sponsor should be an Industry Expert from outside to provide the external practical point of view. The role of the CoE should be multi-pronged. It should do the market research and gauge the direction to which the market is moving. Based on that, the COE should identify the tools and technologies that it needs to procure, prepare training material, and onboard the correct partners as necessary. This body is responsible for the outcome of the efficacy of the training.
- The next step is to make the tools and technologies available to the students. It is common knowledge that Data and Analytics have different tools which can do the number crunching, cleansing, and collection of data. Depending on the maturity of the industry, the students need to get practical knowledge of building and implementing data science models using various tools. Educational institutions should get the right tools on board. While there are some open source tools, others might have a license fee associated. The COE must be involved in making the business case for the procurement of such tools and expedite the process.
- Over and above the tools, educational institutions also need expert facilitators and training materials for which deep connect in the industry is required. This is where Industry leaders and the accomplished alumni come into play. These folks are at the top of their industry, and their extensive experience can act as a guiding force for creating an industry-ready curriculum. Along with reaching out to Industry alumni, the educational institutions can collaborate with industry experts and analysts and draw from their expertise. The industry experts can also help the students catch industry-specific nuances. They can further add value by helping design the content and construct.
- Having a good connect with recruiters and training partners can also go a long way in designing the right training material for the students. The relevance of the course material is of utmost importance. Data science needs three viewpoints – Domain, IT skills, and statistics knowledge are all extremely necessary to ensure employability.
- Last but not least, the students cannot become employable without actually getting their hands dirty in real-world challenges and scenarios. With this as the ask, internship opportunities in leading organizations become crucial. The alumni and connects in the industry can play a vital role to ensure that the future crop of graduates are more employable and data science savvy.
We have seen that industry-academia partnerships and constant feedback could play a crucial role in ensuring the employability of fresh graduates. It is especially necessary for cutting-edge areas such as data science and analytics. There is no silver bullet that can be bitten to solve the problem overnight. Of course, there will be challenges like infrastructure, to build or buy. We are all going to commit mistakes. But accepting that there is a problem and take baby steps now to solve it is what will matter.
This will help institutions and its Students to promote groundbreaking Research & Development and prepare tomorrow’s Data Science Researchers to solving global and complex problems with a scientific approach and capabilities. This works as a key enabler to promote the development of relevant skills and low-cost innovations, promote start-up incubation, entrepreneurship development and better employability.
We at Rubics Labs are working with several leading universities to help them set-up and operate Data Science CoEs by partnering with various tech companies and data savvy people. If you have an interest, drop me a line..