BUSINESS TECHNOLOGY MEDIA COMPANY
Companies
Sectors

How to make data science work for you

Data scientists should have experience in working with volumes of disparate data and an understanding of what a business truly needs from its data.
Read time 4min 10sec

The ‘next big thing’ for a while now has been the emergence of dedicated groups of highly-trained individuals employed to deploy and utilise a very specific set of skills to help them, and by extension business, unify statistics, perform basic and complex data analysis, interpret and design informatics, and their related methods in order to understand and analyse actual phenomena with data. Enter the data scientist.

The Oxford Dictionary defines a data scientist as a person employed to analyse and interpret complex digital data; for example, analysing the usage statistics of a website, especially in order to assist a business in its decision-making.

An evolving science that first hit the mainstream around three years ago, many companies and individuals jumped on the data train and the skill set as a whole became a much-overhyped mismatch, and in my view, hodgepodge of definitions and interpretations. To this day, it is difficult to find one overarching definition that sums up the actuality of what a data scientist does and can bring to the table.

At the end of the day, there is a level of confidence that comes with time.

As such, there is a very real perception that the skill set as a whole has been oversold, and by and large, is not living up to the expected returns. I disagree.

Data scientists are not trained. Yes, there are many courses and degrees and such that can help an aspiring data scientist on their journey, but in my mind, a data scientist is forged. Forged in the years of experience and applying business acumen to enhance decisions.

That sounds almost like something out of an adventure movie, but bear with me…

In my opinion, data scientists need to have a level of experience of working with volumes of disparate data, but also a level of understanding of what a business truly needs from its data and how the discoveries made can be utilised to their full potential.

The courses and degrees that are available for study are commendable and give all aspiring data scientists a good foundation. However, my view is that there is a slight over-emphasis on coding and math, and newly-minted data scientists don’t always know how to apply what they have learnt in a business context. I think there should be an increase in emphasis of business knowledge, improved training in language skills, and data storytelling should become a fundamental part of the journey.

At the end of the day, there is a level of confidence that comes with time. An inherent knowledge that comes with experience and a fundamental understanding of how all the disparate bits of data fit together.

All that being said, and no matter what your definition is of data science, I think we can all agree that the functionality of data science is governed by the readiness of a particular business to be mined. The accuracy, completeness and governance all need to already be in place before anything meaningful can even be hinted at.

We can also agree that data engineering enables artificial intelligence (AI) to a large degree. So, while AI can significantly add business value, if the data engineering portion of the data discovery is not grounded in experience and a fundamental understanding of what business needs – then the fight is already half lost.

Once these models are discovered, designed and developed, after business approval – these models also still have to be moved to production where their true efficacy and success can be established. This is not typically a data scientist’s forte.

However, staying close to the production model and having built a solution that is robust enough to be tweaked and trained on the fly – that absolutely should be part of the fundamental tools that are in the briefcase of data science.

So, while the theory of data science combines multiple fields (statistics, scientific methods, AI, data analysis to extract value from data) it also requires a massive treasure trove of skills. Some can be taught (code, statistics, math, algorithms) – but some of those skills come from experience, the experience of working with data and the experience of working in an ever-growing and changing business environment.

Actionable insight is simply not possible without actually providing to business the answers to questions they need answered. Finding the nugget in the straw, if you will.

However, finding a nugget and being able to sell that nugget to business is what truly matters. The confidence to do that requires both training and experience.

Jessie Rudd

Technical business analyst at PBT Group

Jessie Rudd is a technical business analyst at PBT Group, a position she has held since 2011. In this role, she is responsible for combining data analysis assignments and researching new technologies in this space. Rudd holds training in IT (computer management) and has been exposed to a number of industries over the past 10 years, including BI, financial services, retail, market research, as well as corporate functions such as call centres, human resources and IT. This broad experience allows her to grasp the complexity attached to converting data into intelligence. Rudd has a passion for investigating new technologies and making others aware of them, as well as finding the most efficient tools for successfully undertaking a required task.

See also