Dive into Data: A Career in Data Science
Next up as part of our “Dive into Data” initiative, we aim to change the misconception about what it takes to break into and build a long-term career in the data industry. As part of this, we’d like to feature the background and career journey of Tom Greenwood, Data Scientist at Artesia Consulting Ltd; a brilliant example that showcases the many avenues and routes.
Our chat here, to feature a real life example. How to get into data science and an example of how a career path within that field can unfold….
ADLIB: To summarise, what are your main responsibilities as a Data Scientist?
Tom Greenwood: I have a pretty broad scope at Artesia, which is one of the things I really like about my job.
Inevitably, there’s a lot of data wrangling, which is generally thought of as the less ‘glamorous’ end of data science. However, once you learn about functional coding and pipes and string manipulation, it can be quite fun. I also do a lot of modelling, all the way from simple linear models, to things like random forests and neural networks.
The next thing is visualisation, which also can be as simple as a line plot, all the way through to an interactive multi-part plot in d3.js, or an animation. Once you have an HTML page to host your visualisations in, it really opens up what you can do, so this means either writing apps to present the work, or markdown pages.
We build the apps in Shiny (which is an app framework for R). I really enjoy this part. Designing an app so that it can enable insight into the data is good fun, and app logic is a totally different challenge from pure data analysis.
I also get involved with web deployment issues, like hosting, security, and DevOps. For example, I recently put together automated build for an app using docker containers and Gitlab runners in the cloud. Like most data scientists, I hadn’t even thought about this part of the job until it suddenly became very important. Now, it’s one of my favourite parts of the job.
ADLIB: What’s your background, what has been your career journey so far leading you to where you are right now?
Tom Greenwood: My background has been quite varied. My first job was on the news team at New Energy Finance (NEF), in London. The company was bought out by Bloomberg, which was an interesting experience.
I transferred to one of the analytical teams at NEF, and from there I made a move into carbon markets research at a company called IDEAcarbon. I ended up in charge of the team’s research output and a director of the company. We did a lot of quantitative analysis in excel, and economic research, but no coding.
Whilst I was at IDEAcarbon I became interested in painting, and went to a few classes. Eventually I went part-time at work and started studying at an art-school part time, then full-time and then began teaching at the art-school I’d been studying at.
After teaching art for a year, and trying (not very successfully) to sell a few paintings, I moved to Bristol and set up a new branch of the art school I’d been teaching at. The branch did well enough to break even, but not well enough to pay my salary, so I began to start thinking about another career change.
I’d always harboured an interest in coding, so I started learning online. I also went to lots of tech meetups around Bristol. I really liked the people I met and found the talks interesting, so I started applying for jobs.
ADLIB: What attracted you to data science?
Tom Greenwood: The first thing that attracted me to the area was the breakthroughs being made in image recognition. As an artist I was fascinated by the way computers were learning to recognise objects the same way humans can.
Another motivation was from seeing the huge impact on everyday life that technology and data science is having. I wanted to be a part of this sector which is playing such a big role in how our lives are changing today.
Finally, the overall health and growth prospects of the industry helped as well!
ADLIB: What do you see as the top 3 skills it takes to become a Data Scientist?
- A ‘have a go’ attitude. When you start coding, a lot of stuff doesn’t make sense. There is so much jargon, so many different languages, packages, choices, etc. It can be overwhelming. The best piece of advice I ever received when I was getting into data science, was to just start somewhere.
- Humility. Data science can be frustrating. When a piece of code works, it’s great, but sometimes you can spend hours stuck on a problem. Take a break, go for a walk, read stack-overflow or ask a question online. The programming community is unbelievably altruistic. If you have a problem, guaranteed there will be someone, who’s never met you, willing to help.
- Practice. I don’t know if this is technically a skill or not, but it’s true. You won’t become a data scientist overnight. In programming, skills are built on top of other skills. Do online courses, follow online how-to guides or start a small project. Whatever you do, just make sure you write lots of code and run it, because writing it is ten times better for learning than reading about it.
ADLIB: Any soft skills that are advantageous to this role in your opinion?
I think soft-skills are some of the most important skills for a data scientist. Data science is a team game. If you’re in a company, you won’t be building an app or undertaking a big piece of research on your own; you’ll be doing it as part of a team and that means you’ll need all of the soft skills that are expected of other people in professional teams.
Of course, I’m still fairly new to the career, so I can only talk about my own experience, but so far I’d say that data science is much like a lot of other careers in that people who can listen well, empathise, employ diplomacy and make a client feel valued, will always be in demand. Those skills are just as hard to learn as building a neural network, if not harder, and just as valuable.
Here, you can check out Tom’s coding blog.
Thanks so much for sharing, Tom!