BAI Interview Series: Dean McKeown on Data Science Skills



In the third instalment in the Business Analytics Institute series of Data Science interviews we talk to Dean McKeown. He is the Associate Director of the Analytics Masters program at Queen’s University, Ontario, Canada. His research interests include big data, analytics and governance. We talked with him about Data Science skills and the data culture in business. 

Which technical data science and management skills do you feel a management candidate should have? 

From the outset you need to have a diverse set of skills. There has to be the mathematics and statistics knowledge in order to verify your findings and prove correlation or how significant the statistics might actually be. You have to have business knowledge which means understanding what business you're in, what you're trying to sell or provide or support or consult on because in Data Science, you don't know what you don't know. And that gap in knowledge makes it extremely difficult to do any sort of modeling or data analytics because you may actually be missing a very important variable and it makes you susceptible to biases.

Having an understanding of the technical side of data governance is also extremely important because of you can  address issues of when you can use the data, where do you find it, or is it OK to access that data once you're a behind a company firewall and so forth.

If you're going back to the very basics you should have an understanding of the popular programming languages in the field and have a good grasp of the tools used for example Tableau, SAS. If you lack the skills, you would need to have the ability to learn new tools very quickly, which is more important than actually knowing how to use those tools.

Cognitive skills refer to an understanding of how people interpret the data they see, as well as how they use data to incite action. Which specific skills do you feel a data scientist should have? 

Cognitive skills are about understanding the business that you're in figuring out what goes on in there. A company cannot achieve success on data analytics alone, they have to understand their goals. So you have to be able to develop a strategic outlook of what's going on with in the business and establishing goals.

A data scientist uses scientific methods of developing a hypothesis and testing, gathering the data and learning and adapting the hypothesis and then going through that continuous process of testing and learning. This requires critical thinking skills. And you can't develop these cognitive skills without having a very strong understanding of the industry that you're in, the products that you're offering or customer base that you want to target.

Empathy is an important piece in this puzzle because it allows you to understand the customer's viewpoint to analyze what they really need and how your company can help solve the customer's problem.

Presentation skills play huge part here too. You should be able to take your data and analytics and translate them into a presentation or a written document that people in the business can understand.

In your opinion, for what should organisations or HR agencies be looking in recruiting the right people for data science roles?

Certifications are very useful for the assessment of technical  competencies. You can get certified in SAS. It's more difficult for the open source tools because they haven't quite got that community who can certify you though IBM is currently working on such certifications.

Companies need to find people the are able to learn tools quickly as well as demonstrate critical thinking and soft skills like visual communication which are going to be so important in translating  the numbers into insightful decisions.The person that they are looking for should have interpersonal skills to get along with the team and leadership skills to be able to steer collective decisions.

It's very important for  companies to understand exactly what they're looking for. There are really some very distinct roles within the moniker of data scientists, for example you will have database administrators, data cleaners, business people who are data savvy, developers and so on. And so the company has to have a very clear definition of what they are looking for. HR  people have to get a handle on what's going on within business as well work very closely with the business unit head to understand what specific skills are being sought after . They have to be clear in what they're looking for and making sure that the people that they're actually interviewing and hiring meet these very specific criteria.

Can you offer examples of companies that have a particularly good data culture?

It’s difficult to give to pinpoint a company as a whole however there are pockets within most organisations that get it but there's still a lot that don't. A lot of companies don't trust the data therefore they can't use it to make decisions and they fall back on the gut instinct all the time as opposed to looking at the actual numbers -but this introduces personal biases. The most successful companies are the new tech companies like Google and Amazon. From the very beginning, they had a rich data culture about data  and that's why I think that they have become successful so quickly. Amazon is a great example because their goal is to become the one stop shop for everything from toothpaste to high end television screens. They have built loyalty so that customers trust them with their data.

They have a huge advantage because right from the start they were data-centric and their data works well together. Whereas the more traditional companies more often have delocalised data centres and then each business unit has its own databases. You may even have individual business units who will buy data sets that are totally different than what the main company has and then they try to bring all these data together. That sort of diversity creates a kind of complexity that is difficult to overcome and it doesn't help the company.Constrastingly, companies like Amazon started with a couple of simple data databases and then built up from there but they always had one core fundamental piece with which they could connect.

What advice would you give to to a person that wants to pursue data science as a  career?

If you want to become a data scientist you have to carefully consider your goals. As a data scientist you'll most likely be be slotted up in a corner to crank out data. The data and your analysis will most likely go to business person who wants to be a director or CEO. And they will utilise your analyses to make decisions and incite action. It's therefore very important that you go into these roles understanding exactly what you hope to achieve.

In the next 5 to 10 years, there will be a reorganisation where you will have  data groups that look after the data- for instance cleaning the data, establishing the governance of that data and then that group will work very closely with the database administrators or I.T. groups who  will ultimately provide reports to the business units in marketing, sales, finance. You can't go into the world of data science thinking that you're going to grow into a CEO at some point in time. What's important is, if you want to be  that successful CEO in the future you have to have a very good understanding of data analytics today.

And this is one of the issues today in the Data Science world, there is a knowledge gap between management and data analytics.We run a Masters of management analytics program and that's exactly the niche spot that it fills. It is a business degree with a heavy emphasis on analytics as well as understanding how to do evidence based decision making by analysing data. It also teaches students how to communicate with data scientists and CEOs, who each have a unique way of communicating. You have to be that bridge between the two.

Dean McKeown is an Associate Director for the ScotiaBank Centre, Customer Analytics at the Smith School of Business, Queens University Kingston Ontario, Canada. His LinkedIn profile can be viewed at This is part of a series of interviews by the Business Analytics Institute  on Data Science and Data Science skills.