A Career Conversation with Andrew Howe, Wood Mackenzie Director of Data Science
Recently, Andrew sat down and shared his thoughts and experiences throughout his career in data science. From offering advice to new starters to giving an insight into what he is looking for in a successful data scientist, find out what he had to say in this career conversation.
What would you do if you were on your first internship?
As an intern, no matter what your job is, you really have three jobs for your career development:
- Judge if the career path you are currently following is right for you. This is of paramount importance. Is it interesting and rewarding? Do you have the mindset, talents and skills to be impactful and successful?
- Make it a point to learn something from as many colleagues as you can. This is clearly true vis-à-vis seniors on your career path but is just as true regarding people in other departments or domains.
- Strive to maximise your exposure to a variety of technologies, techniques, and domains. Thinking as a data scientist, how about exploring statistical simulation, operations research, mathematical optimisation, or agent-based modeling? You will always learn something, often something that expands your horizons and contributes to the development of your career.
What is the best advice you can give to help plan a career, rather than simply working to keep a job?
Abandon the concept of a linear career path and adopt a flexible career portfolio mindset. A career in the modern workplace is less about carefully planned steps taken in a particular order, and more about flexibility and agility. You should be able to adapt to the changing landscape and capitalise on your experience to take advantage of new opportunities. If you’ve been successful with the three jobs in your internship, this is easier. You will have a good idea of what is typically required for your chosen career, how you meet those requirements and how to start developing career agility.
Put together a career development plan which should direct your career in the direction of where you would like to work or what you would like to do. To whatever extent your skills are lacking, your career development plan should include their acquisition. Since this is a long-term plan, there are two important factors to keep in mind:
- Development of a career portfolio should always be flexible. Rigidity in following your plan will lead you to miss deviations which could give valuable experience.
- When you are considering a job move, you should assess the extent to which it moves you closer to your ultimate career goal and enhances your career portfolio. In evaluating a potential job move, don’t let compensation or job title be your primary considerations.
Finally, career development is less about job titles; it’s about opportunities for growth through multiple experiences. These experiences form the basis of your career portfolio.
What do you think about having a personal brand; what is yours?
A concise, explicit and specific personal brand is a great idea. We all have a personal brand — it’s what others say about us when we’re not around. A personal brand is a useful tool for selling yourself to others and differentiating yourself from your peers. It also helps focus your own thinking about yourself and your career. But it can take concentrated thought and effort to develop one. Personal brands can include several components. Along with my CV, I have a single-slide summary of my qualifications and accomplishments in my current job. I also have a three-slide detailed presentation of my brand. It’s critical that a personal brand includes a two-minute elevator speech.
The elevator speech is what you would tell a CXO during a short elevator ride. It is intended to pique their interest in what you can bring to the table. While clearly about you, an elevator speech should not focus on you, but should focus on your audience: your goal is to share what you can do that will solve a problem of his. My elevator speech introduces me as a “quantitative modeler who uses mathematics and computer science to solve impactful and challenging business problems”. Notice how I immediately address why he should care, and how I differentiate myself from my peers (quantitative modeler vs. data scientist).
How can you be innovative?
Innovation and creativity are not about coming up with one good idea and sticking with it.
Creativity and innovation are more directly related to coming up with a lot of ideas and throwing them away to get to a better one. A person will often generate a good idea that may have clear flaws and only focus their efforts on fixing the flaws, rather than moving on to the next idea — which may be great. This tendency kills innovation: resist the urge. Iterate on ideas, never being afraid to throw them away; in this perspective, innovation is not much more than a numbers game. Once you have that great idea, only then should you grasp it firmly and iterate to improve it.
How do you continue learning while working full-time?
I consider continued personal development to be critical to my career success and progression. Continued development is especially important given my position as a technical leader. I try to always be learning something. I learn by reading academic papers, taking online tutorials, Udemy/Coursera courses, and more. I also study topics without an immediate business need, as long as I think they may be useful later on in my career.
I expand, develop, and hone my skills in several ways. The first way is I block time during the week to do some learning on my own. This is difficult to do and requires discipline and a big-picture mindset. I also tend to learn a lot on the job. For example, I learned data pipeline orchestration with Apache Airflow through an online course but didn’t “learn” it until I needed to apply the technology to a project. Finally, what I do as a quantitative modeler is something I love doing and one of my hobbies. I’ll frequently read about a new algorithm or technique and want to explore it more so I spend some of my personal time learning, coding, and exploring it. For instance, I’m interested in genetic algorithms, and also baseball. One of my recent projects has combined these to apply genetic trees to historical athlete performance data to attempt to engineer a single feature that predicts future team performance.
What advantages do your Ph.D. and MBA give you as a data scientist?
My higher education has been very helpful in my career. The knowledge and skills I gained are directly applicable to my daily job. Nearly all the modeling algorithms in data science are founded in statistical science. My Ph.D. in statistics gave me a solid foundation in quantitative modeling. Considering my MBA, I believe practical business domain knowledge is essential in data science. Familiarity with the business domain leads to a better understanding of specific requirements and perspectives on how to solve impactful and challenging business problems. For example, a model may find something that is statistically significant. However, it may be significant but not sufficiently substantial to warrant the expenditure of limited resources to act.
For students in STEM fields, an MBA is often an afterthought, and not everyone can pause or delay their career to get an MBA full-time. If this is the case, take a part-time degree course/classes, or make an extra effort to maximise your exposure to practical business thinking on the job. This will help expand your career portfolio. Another benefit of higher degrees is that they signal the marketplace about qualifications. Higher degrees in a core topic signal depth of familiarity. Degrees in potentially related topics signal breadth of perspective. Degree signaling has helped with my career progression substantially.
What are the most important skills you look for when hiring data scientists?
This depends greatly on the seniority of the role I’m filling. With a senior role, I place more emphasis on soft skills such as leadership, project management, and big-picture thinking. That said, our jobs are based on programming computers to acquire data and perform mathematical analysis of it so there are several foundational skills and character traits I seek. Technically, I want someone with strong statistical and SQL skills. As I stated earlier, statistics is the foundation of most data science.
In addition, no matter what latest advanced database systems come out, I think the majority of data will always be stored relationally and queried by SQL. Stronger SQL skills allow more computation to be performed on the database server, increasing computational efficiency. Sufficient proficiency in a relevant programming language is also vital. These are foundational technical skills.
Finally, there are four important soft skills I try to find: communication, collaboration, creativity, and curiosity. In my opinion, these four Cs are vital for a successful and impactful data science career.
