Data Science for Decision Makers: A Discussion with Dr Stelios Kampakis
This article contains an interview veteran data scientist, Dr Stylianos (Stelios) Kampakis, in which he discusses his career, and how he helps decision makers across a range of businesses understand how data science can benefit them.
By Jo Stichbury, Freelance Technical Writer
In this article, I’m interviewing a veteran data scientist, Dr Stylianos (Stelios) Kampakis, about his career to date and how he helps decision makers across a range of businesses understand how data science can benefit them.
While data science is a field showing immense growth at present, it’s somewhat nebulous in its description. I think there’s a lot of uncertainty as to exactly what it is and how to apply it. Fortunately, Stelios is an expert data scientist with a mission to educate the public about the power of data science and AI. He is a member of the Royal Statistical Society, honorary research fellow at the UCL Centre for Blockchain Technologies and CEO of The Tesseract Academy. A natural polymath, with a PhD in Machine Learning and degrees in Artificial Intelligence, Statistics, Psychology, and Economics he loves using his broad skillset to solve difficult problems and help companies improve their efficiency.
Data Science is an increasingly popular career choice these days. Stelios, how did you come to work in the field, and how long have you been doing it?
Well, that’s an interesting question. Many people these days end up being data scientists solely because there are great job prospects. You see people having studied something completely different (e.g. chemistry), and then they educate themselves in data science using various online materials.
My case was quite different. My first degree was in Cognitive Psychology, and I was planning to become an academic in the area of cognitive science, AI and neuroscience. I ended up spending more and more time in data analysis, until I decided to do an applied PhD in machine learning. This was the transition point for me, as I moved away from academia and into the industry. However, I would say that my involvement with the wider area of data and AI, has lasted more than 10 years, and started in my early university days.
Of course, data science is a broad umbrella term. How would you break the field down into different subject areas? Do you have any preferences for any of them?
That’s a great question, because not many people understand that. Before the term data science we had people working in distinct fields such as rules-based AI, machine learning, statistics, computational intelligence, etc.
Data science as a term means “I am doing stuff with data”, and it is more of a term used to make it easier to sell these kinds of services, because understanding the subtle differences between the various schools of thought, is way too confusing for the uninitiated.
In my book (“The Decision Maker’s Handbook to Data Science”) and workshop I explain that there are three main fields that have shaped data science: classic AI, machine learning and statistics. Classic AI is not so much around anymore in terms of techniques, but it has played a huge role in shaping part of the history of modern AI.
I am one of the few people that have received training in all of these areas (and some more, e.g. computational intelligence). I don’t have any particular preferences, and I’ve done work using methods from all of them. However, if I had the possibility, I believe that the most fascinating opportunity would be to work on general AI.
What is your advice for anyone wanting to get into data science?
Hm, I think that getting into data science now is the easiest it has ever been. I know some people will disagree with me. I also know, that me and other people have spent countless hours in our education and training before we called ourselves data scientists, so it might not seem fair to say this. However, let me explain.
I believe that there are three main layers of data science expertise.
The expert: You have a wide coverage of most areas of data science, and can do research in the field, or propose new methods. This usually describes people with 7+ years of academic experience alongside a few years of industrial experience.
The practitioner: You have some knowledge of how techniques work, know how to code and how to use some tools, but your knowledge is not very extensive. This describes people with some academic and industrial experience (e.g. an MSc degree).
The tool user: You are familiar with some basic techniques, and know how to use some basic tools and libraries. These people might come from different fields (e.g. physics), and they may be good with using tools, but they may have somewhat limited coverage or understanding of some of the areas of data science.
I think that most people that self-educate themselves are between the practitioner and the tool user archetypes. And in reality, the majority of jobs can be done by these people, as they don’t require an excellent understanding of how methods work.
We had seen a similar trend with software development. You had to really know about computer science in the past, if you were to code the simplest thing. Now, many software projects can be accomplished by people who have just spent a few months going through tutorials and using high level frameworks and platforms.
This is not to say, however, that experts are not needed. They will just cover the upper end of the market.
Do you have any thoughts on where we are heading in the future? What will the next 5 years look like?
I believe we are going to see more of the same trends. Data science will expand across more and more verticals and more and more countries. I’ve been doing work with companies from all over the world (from Egypt and Cyprus, to the US and Germany). Industries in all countries are trying to catch up with the trend, collect more data and make better use of it.
We will keep seeing a shortage of data scientists, but it might not be as bad, given that (as I mentioned in the previous question), we will see the appearance of many data scientists with limited skills, who will be sufficiently good to solve most problems.
Can you tell us about your upcoming event that covers data science and AI?
It’s coming up soon, on March 21st 2019, and is a workshop that covers everything a decision maker needs to know about data science, such as how to use data science in a company, or how to hire and manage data scientists.
The motto of my company, The Tesseract Academy, is “technology made simple”. Most of the resources and events out there are either very technical, or very fluffy. The goal of the Tesseract Academy is to deliver training in this fine balance, that helps a decision maker understand what data science is about, and how it can be used, but without assuming that they will develop anything themselves.
I have people from all backgrounds coming to the workshop. Investors (who want to learn more about the subject), recruiters (who want to understand more about how data scientists think), CEOs of scale-ups or startups and product managers.
I think you recently published a book — what is it about?
It is the “Decision Maker’s Handbook to Data Science”. Much like the workshop, it is a high level overview of data science, and covers everything a decision maker needs to know about data science. It includes many case studies from my experience, as well as various examples from multiple industries. I also hand it over for free to the participants of the data science workshop
It is available in PDF format from my blog, but also on all major e-book platforms (Kindle, PlayStore, iTunes, etc.)
Finally, besides the resources you have published, what are your recommended go-to websites, books and courses?
I guess this depends on what path of your learning journey you are in! I am working on releasing a new training program for aspiring data scientists called Datalyst. However, I, myself, am concerned with keeping up with the latest trends on research and the most cutting edge techniques.
I am reading the blogs of all major companies doing research in AI (Uber, Google, Facebook, etc.). I am also reading the papers from all major conferences (ICML, NeurIPS, AI & Stats, etc.). Finally, there are some blogs and newsletters I am following, such as Data Elixir and Towards Data Science.
Thanks Stelios, it was good to get some clarity from someone so clearly gifted with the ability of simple explanation, which is close to my heart! I’ll look forward to find out more from your workshop in March, and wish you — and all budding data scientists — happy number crunching!
Bio: Jo Stichbury is a senior software professional with 20 years experience and Freelance Technical Writer with interest in AI, data science, AR & VR, science and science education, software engineering and agile/lean development
Original. Reposted with permission.
- Deep Conversations: Lisha Li, Principal at Amplify Partners
- Don’t learn Machine Learning in 24 hours
- Becoming a Data Scientist