Improve the value you can deliver by developing soft skills.
Data science is a rapidly growing field with the demand for data scientists projected to grow by 36% over the next decade according to the US Bureau of Labor Statistics. However, the field can be very competitive, and there are many steps between being an entry-level data scientist and getting promoted into a senior role at a large company.
Even though it is great to have sound technical skills within data science, soft skills are equally, if not more important to develop. Developing these skills will help you stand out from the crowd, and also serve you in future roles if you decide to change your career path or move onto more senior roles.
Within this article we will cover 5 essential skills that you should develop in order to be a successful data scientist.
Communication Skills for Data Scientists
Communication, verbal or written, is essential for data scientists. It’s not just important for the quality of your work, but also for how you interact with your team, clients, and senior executives.
Developing good communication skills will help you to share complex ideas clearly — even if that means simplifying things so everyone can understand them easily. This will make it easier for you to explain your findings, which is especially important if you’re working with a client who may have different needs than yours or has never seen similar work before!
Additionally, if your role involves training junior members of the team you may need to simplify complex ideas and topics into something easy for them to understand.
When carrying out oral presentations we can sometimes use presentation slides as a crutch. This can detract from the overall message and delivery. Instead of relying on slides 100%, they should be used to show extra information and highlight key points to enhance the audience’s understanding.
When I started in petrophysics, which involves dealing with a lot of data and delivering reservoir studies to clients, my presentation skills were dire. I would often present more information than was required. After several years, I have improved significantly and have delivered multiple presentations, including at conferences.
If you want to improve your presentation slides, I recommend checking out the following video:
Domain Knowledge for Data Scientists
Many people unfamiliar with machine learning and data science assume that you feed data into an algorithm, click a button and as if by magic, out comes the correct answer. However, the reality is often quite different.
Having knowledge of the domain you are working in helps you understand the problem and its relevance to the business goals. Without this, it can be difficult to identify the correct methods to apply to the problems, and whether the results generated by your algorithms make sense.
Within the petrophysics or well logging domain, we take measurements of the subsurface using a number of scientific tools. These include recording the formation’s natural radioactivity and response to electrical currents. Sometimes, for a variety of reasons, we may get responses that we are not expecting or are the result of physics. These odd responses may be inadvertently removed or go unnoticed and could have an impact on the final modelling solution.
If you are intending to stay within your role for the long term, then it may be wise to learn the basics behind the domain you are working in. This way you will have a better understanding of the data and will be able to communicate more effectively with other stakeholders.
Alternatively, if you are working in a team it may be prudent to discuss the background of the data as well as the results with a subject matter expert.
Business Acumen for Data Science
Having a sound understanding of business is essential to progressing far as a data scientist. It allows you to understand how different segments and processes within a business work, as well as being able to understand the problems that a company may face. Once you have an understanding of what a company is trying to achieve, it will be much easier to translate the business problem, select the most appropriate solution and connect it to business impact.
When working as a data scientist it may be tempting to build a large and complex model that you think may solve the company’s problems, however, it may be the case that a simple data analysis would have been sufficient.
Sometimes simple solutions provide the most impact.
For example, carrying out a funnel analysis of how many people have abandoned their shopping cart during checkout can help the marketing team focus their automated email campaigns to improve conversion rates.
Teamwork Skills for Data Science
Working with others is one of the most important skills for data scientists, and it’s easy to see why: teamwork is key to success in many fields from software engineering to medicine and even law enforcement. A good team will help you get more out of your workday than an individual effort could ever hope to achieve.
When working on larger projects or within a company there will be a requirement for people with different levels of expertise and capabilities, at different stages of the project.
For example, if a project involves creating a commercial product. In the early stage, there may just be a requirement for data analysts, subject matter experts and product managers. Once the project starts to gain momentum, other disciplines such as software developers and machine learning engineers may be required.
Columbia University put together a series of tips on successful teamwork within Data Science that is well worth a read:
Problem Solving Skills for Data Science
This skill is essential in any field where data or information is used. A good problem solver can identify problems quickly, come up with solutions quickly (if possible), implement those solutions efficiently and effectively and communicate their progress regularly with stakeholders.
When starting a data science project the first step before any work is carried out is to define the problem statement.
At first, the problem statement may be high-level and vague, for example: “We need to improve our profits”.
But with some tweaking and investigation, we may end up with a more refined and focused problem statement such as: “What are the most effective processes we need to implement over the next 6 months to increase profits by 10%?”
This now provides us with something that is clear, measurable and precise that will allow us to come up with the most effective and efficient solution(s). It also saves wasting time from going down the wrong path.
The following article goes into more depth on how to refine the problem statement and is well worth a read.
When learning data science the majority of the focus is on developing technical skills such as data visualisation, python, R or machine learning, but soft skills should not be forgotten or neglected.
Not only will developing soft skills such as communication, problem-solving and business acumen help you in your current role, but they are also transferrable and can influence any future roles you may go for.