Non-Technical Principles All Data Scientists Should Have
Making you a better data scientist, and enhancing your career.
Introduction
There is a common misconception that the data scientists who are most likely to succeed are the ones who have the most technical ability. I have been working as a data scientist for almost a decade, and during that period I have been promoted multiple times. Although I do have technical skills, I also have to attribute my success to my non-technical principles.
I don’t agree there is a strong correlation between technical ability and career success, instead, I would rephrase this statement as:
Data scientists are less likely to progress in their data science careers if they do not have/develop their non-technical principles.
Success can be defined in various ways; this could be financial, working for a specific employer, having a good work/life balance, or having a specific job title. Regardless of how you define it, I still believe this statement holds.
The purpose of this article is to discuss key non-technical principles all data scientists should develop to enhance their careers. Principles will be divided into either an attribute (personal characteristic) or a habit (routine behaviour).
Attributes
An individual's attributes are often inherent or learned over time, e.g. being able to adapt to a changing environment. Although difficult at times, attributes can be taught.
Whenever I am interviewing candidates for a data scientist position, I always make sure to evaluate their non-technical attributes as well as their technical ability. I find technical skills are much easier to teach compared to non-technical attributes. Demonstrating these during your interview can really enhance your chances of being hired, it is not always about being able to explain how a neural network works!
Soft Skills
Soft skills is a broad term, so let's break it down into individual traits.
Communication
Often overlooked, having good communication is not only about being able to speak well, it is also about your ability to listen. As a data scientist, your audience is mostly non-technical, a key skill to develop is the ability to explain highly technical concepts to less technical audiences.
Something I learned early in my career is to “always know who your audience is”. For some audiences, you need to avoid technical terms and place more emphasis on the “why” and “how” of your results. Other audiences might be more interested in understanding model explainability, therefore you should tailor your presentation pack to include more technical content.
Communication can be developed with practice and experience. Whether in-person speaking or via email, working on your communication skills will help you build better relationships and enable more people to understand what you are working on.
Collaboration
Data scientists are positioned within an ecosystem that demands collaboration to be successful. Not only do you need to collaborate within your data science function, but you also need to speak to external teams e.g. software developers, data analysts, machine learning engineers, product managers, and stakeholders.
I highly encourage all data scientists to participate in group discussions and to make sure you’re actively challenging each other, regardless of job titles. A great data science team is one that supports each other on their tasks, as well as offers constructive criticism to help aid personal development.
If you’re not doing so already, reach out to people in other teams who are contributing to the same project, send them a virtual 1:2:1, or have lunch together and introduce yourself. I’ve found that developing non-work-related relationships with people actually benefits me when I need to speak to them about something work-related.
Problem-Solving
The primary role of a data scientist is to take a business problem and translate that into something that can be solved using data. As much as we all love to dive into training machine learning models, there are some fundamental steps you need to complete prior.
My first piece of advice which I believe is a must-learn, is ensuring you fully understand what the business problem is before you go diving into the data. Speak with stakeholders, ask questions, and make sure you fully understand what the problem is. Remember, stakeholders and directors are not as close to the data as you, you are the expert and should be using your knowledge to collaboratively come up with the best approach to solve the problem at hand.
Once you understand the problem, break it down into smaller tasks and manage these systematically. If your team works with planning software such as Jira, this will help as you more than likely need to perform this step anyway. Not only does this allow you to identify what subtasks are dependent on other tasks being completed first, but it also allows you to delegate subtasks to other members of your team should priority increase.
To be good at problem-solving, you must be creative and have the ability to make the correct decision, often on a time constraint. I always encourage data scientists to be innovative in their thinking, there is not one way to approach a single problem. There will be instances where deadlines are coming up and the room for exploration and innovation shrinks, but in instances where this is not the case, I highly encourage people to think outside the box.
Business Acumen
A data scientist’s role goes beyond developing machine learning models. Whatever task you’re working on, you should clearly understand what business problem you’re solving and the impact/value your work will deliver.
I strongly recommend spending time educating yourself on the sector you’re working in. For example, if you work in logistics, educate yourself on how the supply chain operates and the various processes involved. Not only will this enhance your domain knowledge, but it will also help when communicating to non-technical audiences who have spent a significant amount of their careers working in this sector.
Some people naturally display strong business acumen, whereas others require more coaching. When looking to develop your business acumen, I recommend starting with familiarising yourself with the key metrics your team is being measured on. The majority of companies create key performance indicators (KPIs) as a way of tracking progress. Spend time understanding these KPIs, and link the work you’re doing back to how it is going to influence these metrics.
Note: Whenever you’re working on a task, ask yourself, how is this going to contribute to maximising ROI and increasing performance metrics.
Habits
Unlike an attribute which is often an inherent behaviour, a habit is defined as a routine behaviour that over time becomes subconscious, e.g. checking your emails each morning.
Your habits will not be focussed on too much during the data scientist interview process. They will most likely come up during the early telephone rounds as a way of getting to know the candidate a bit more.
Habits are a lot easier to teach, and you will often find when starting at a new company or team you develop new habits subconsciously, that might be how you deliver your updates during stand-ups, how you format your code, or how you write your merge request descriptions.
With that said, there are a few habits that have helped me throughout my career that I wanted to share with you.
Networking
Sometimes opportunities arise due to who you know, not what you know. Working as a data scientist can become quite isolated at times, and the only like-minded people you speak to are the people you work with. I believe that networking outside your day-to-day circle is a non-negotiable.
I first started attending tech-related meetups in 2016, and not only did this allow me to meet more like-minded people, but it also educated me on how other companies’ data science functions work. It was actually at one of my first meetups I met the individual who would end up giving me my first data scientist position. Being able to speak in an informal setting with zero pressure, I was able to showcase my knowledge and passion for data science. If it wasn’t for that meetup who knows how my career would have turned out.
Note: If you can’t attend meetups in person, look for virtual opportunities. You can also connect with people on socials (e.g. LinkedIn) and grow your network from there.
Social media is a great tool when used correctly. During the 2020 pandemic, I began networking with other developers and data scientists on Instagram and Twitter (now X). By doing so, I was able to build connections around the world with developers I still engage with almost 5 years later.
When posting on social media you often find yourself with the mindset of wanting to gain as many followers as possible. When using social media for networking purposes, I encourage you to focus on growing your meaningful connections instead. Meaningful connections are those who follow you back, engage in your content, speak to you via DMs, and share common career goals.
Finding a Mentor
The best person to take advice from is somebody who has accumulated that experience already. Having a mentor does not need to be a formal agreement between two people, simply building a relationship where you can ask questions is all you need.
Often senior data scientists will inherit becoming mentors for junior/mid-level data scientists, the role of a mentor is to provide guidance and help assist people to advance in their current positions.
Personally, I have had many mentors. When I first started working as a data scientist I was mentored by a senior who educated me on coding best practices, and how to become a more well-rounded data scientist. I also had a non-technical mentor who had decades of experience in the sector I was working in, as I was new, we spoke a lot about how the industry works and how data science can help improve day-to-day processes.
Note: Don’t be afraid to reach out to people and ask them if they would be willing to mentor you. Identify those whose career paths correlate with your ambitions, or they have extensive experience in your sector. Reach out to them!
Build In Public
The term “Build in Public” refers to sharing your work publically online. Whether that be creating a portfolio website, populating your GitHub profile, writing articles on Medium, or posting on social media.
I highly encourage others to educate themselves publically. By doing so you’re able to demonstrate your expertise which can potentially set you apart in what is a competitive job market.
I developed my portfolio via GitHub Pages during the 2020 pandemic to expand my knowledge of web development during lockdown. GitHub Pages is a free approach to creating a centralised location for all your experience and online contributions, think of it as an interactive CV.
Linking back to improving your soft skills, I advise all data scientists to explore the world of open-source as early as possible. If you lack experience communicating with other data professionals, working on solving various types of problems, and collaborating with others, open-source could be the thing for you!
If you’re considering diving into developer communities, don’t hesitate. Start by sharing your progress, engaging authentically, and being open to learning from others.
Don’t Compare Yourself to Others
This is a big one, especially when you are starting out in your career. It is important to remember that everyone’s journey is unique, and comparing yourself to others can often hinder your own growth and development.
Data science is an ever-expanding subject. Research labs, universities, and open-source are contributing daily to the already existing large-scale knowledge available. It can sometimes feel overwhelming to keep up, sometimes resulting in imposter syndrome kicking in.
I encourage you to see other data scientists' skills as an inspiration, rather than a benchmark you need to match. If you find yourself in a situation where a colleague presents their work and you feel anxious that you could not have produced that yourself, reach out and speak to them, learn from them.
Every data scientist has their own journey and set of strengths, a good team complements each other. There are always going to be people who are better at certain things. Focus on celebrating your own achievements, avoid being a perfectionist, and recognise mistakes/not knowing certain things are all part of the learning process. Everybody learns differently and at different speeds, there is no correct approach to learning.
Ask for Help
Over recent years I have had the honor of mentoring numerous data science internships. These internships last roughly 6–8 weeks which interns not only have to understand the business problem but also the company data, tech stack, identify the subject matter experts, and most importantly solve the problem itself!
Something that became apparent early was that interns who asked for help had a more successful internship and in most cases landed a permanent role as a data scientist.
Note: What you want to avoid is never asking for help or always asking for help.
Asking for help does not make you look like you’re incapable of completing the task yourself, it is actually a good sign that you’re proactively reaching out to others to help advance your project. Once you realise this is not a negative thing, you will gain more experience collaborating with others, your learning will accelerate, you will be more productive with your time, and the output quality of your project will be far better.
Final Thoughts
Although attributes are more difficult to develop, I strongly encourage you to evaluate your current level and then create a plan to develop these further. Interviewers are on the lookout for these attributes, therefore they should be prioritised when actively exploring the job market.
Your soft skills will develop with experience as well as by adopting habits such as networking, finding a mentor, and asking for help. Developing your business acumen is a must if you’re looking to move into more senior/management roles. You must understand both company and team KPIs, why certain work is prioritised higher than other work, and what value you’re going to generate from the work you’re doing.
Habits can be started today, whether you’re looking to land your first data scientist position or looking for a promotion. If you’re reading this and you do not have an existing professional LinkedIn network, your first task is to create one. Alongside networking, building in public is great for showcasing your skills and unlocking the potential for future career opportunities, I highly recommend spending time exploring this.
Finally, the two more personal habits I believe all people should have are to not compare yourself to others and to never hesitate to ask for help. Imposture syndrome is something everybody experiences, this stems from comparing yourself to those around you. What you don’t know is that those around you all feel the same, and in their eyes, you’re the expert!
Disclaimer: Unless stated otherwise, the author owns all images included in this article.
If you enjoyed reading this article, please follow me on Medium, X, and GitHub for similar content relating to Data Science, Artificial Intelligence, and Engineering.
Happy learning! 🚀
Non-Technical Principles All Data Scientists Should Have was originally published in Towards Data Science on Medium, where people are continuing the conversation by highlighting and responding to this story.
from Datascience in Towards Data Science on Medium https://ift.tt/d2RIbDa
via IFTTT
