How Soft Skills Leverage a Data Practitioner
Without the right mix of passion and talent, data science can be a tough puzzle to crack. Aspiring data scientists have to comprehend the fact that they can’t forge a strong career foundation based on technical skills alone. In 2021, Google conducted an internal study to explore the most innovative and productive groups within the company. Turned out, their best teams were not actually the ones filled with top scientists but those interdisciplinary groups who demonstrate strong soft skills in collaborations.
Data science is a human pursuit. Apart from mastering technical skills, data scientists must polish their soft skills and strike a perfect balance between them. So how do you acquire soft skills? Fret not. Today, we have prepared five of the most common soft skills for data scientists. We will also explain why they are important and provide guides on how to leverage them.
- Communication and Storytelling
Communication, Data Storytelling, or to some Data Politics, is the pillar of almost every type of job out there in the market. This is why it is on the top of our list. This is especially true with data scientists since you have to frequently discuss complex terms with a non-technical audience. For example in customer churns (or simply customers leaving your business), you must be able to explain your model to your team else it won’t be adopted into production. If you couldn’t explain why this model was going to predict customer churn more accurately than the status quo, then no change will be deemed needed.
Storytelling/Data Politics on the other hand focuses on the data practitioner’s ability to persuade others into your recommendations. It allows decision-makers to see your perspective. May it be a compelling data visualization or a catchy storytelling approach, this helps build a strong data narrative to convince your audience.
But can this skill be learned? Definitely! One good practice to adopt is to always make your slide decks readily understood to the point where even the preempted questions could be deduced from your slides. Another step is to simulate your presentation through mock discussions of your slides with your peers. These will develop your communication skills.
As for the storytelling aspect, always design your dashboard according to your target audience. Find out the domain knowledge and technical skills level of your target audience beforehand. You should also always get constant feedback from your supervisor and colleagues on the insightfulness, complexity, and aesthetics of your deck. It’s an iterative process, so be patient! When in doubt, ask this question to yourself: “Does this chart/visualization answer best the business questions presented to me”?
2. Problem-Solving
There is a reason why companies value the experience of job applicants — familiarity with dealing with problems makes you solve them more efficiently the next time you encounter them, which technical education alone couldn’t give you.
There are some ways for you to develop your problem-solving skills. Websites like Kaggle can be a great starting point for beginners. Winners of past competitions usually have good posts explaining how they broke down complex problems into small ones and solve them efficiently there. You can learn a lot from reading other people’s codes as you will learn to think from a different perspective. Additionally, you could also do interesting side projects during your free time and upload them to GitHub. Since you have no deadline for these projects, you can keep improving your solution and ask for feedback from your colleagues or even in online discussion forums.
3. Business acumen
The real value of data practitioners is not on what tools he use, but on what impacts he could derive from using such, in helping an organization to solve its business problems. This is where data practitioners often fall short when it comes to applying theory to real-world problems, especially the business side of it. Although no one readily starts with a business mindset when they began their careers. It is a skill that can be learned through exposure.
For a simple start, surround yourself with people who are more knowledgeable than you about the business. These are usually also the people who would be generous enough to answer your questions or share their insights on how they would do the task at hand in their approach.
Another tip to note — the best way to gain an actual business experience is working within the business itself. Securing an internship at a business firm will transform your life as you get to experience how and why a particular business decision is made. Maximize your onboarding orientations and make sure to familiarize yourself with the domain knowledge and business model of the company. You should also clear your doubts with your supervisors during mentoring sessions. Because ultimately, it is only when you understand the business can you make sense of the numbers you crunch from your tools, right?
4. Critical thinking
Critical thinking in data science can be divided into two parts. First is the ability to ask the right question. The quality of answers you will get relies on the quality of questions you ask. This involves extensive discussion with the business team and stakeholders in order to truly grasp the business problems that need to be addressed.
Next is the ability to question the data. Experienced analysts do not ever dump raw data into a machine learning model and see what comes out because they are aware of the phrase, “rubbish in, rubbish out”. Through experience and sufficient domain knowledge, you will be able to remove potentially distracting or misleading features in the data and add new meaningful features. Usually, this can help make or break a project in an enterprise setting.
The good news is that thinking critically is a learned skill, which can be cultivated by anyone. Critical thinking in a sense is your ability to weigh and act on the most ideal solution. One way to hone it is by never starting a project without understanding the business objective and context. Align at the beginning with your stakeholders what are their expected outputs, their expectations, their preferred format (i.e. tools they are knowledgeable in, design preference), etc… It’s also about being critical of your techniques and algorithms. An example of a question you could ask yourself is whether a problem requires machine learning or not. Items that can be addressed with traditional statistics or data visualization techniques need not be complicated. It’s very easy to apply machine learning to every single problem, but in reality, only a few problems qualify for a machine learning solution.
5. Project management
We don’t call data science projects “projects” for no reason. And in order to be effective and efficient in managing them, you do need to manage the constraints and requirements associated with it. You need to have a timeline, scope, and requirements that you have gathered for your data analysis project as well as resources whether that’s you or other people working on that project.
The further you progress in your career as a data scientist, the better you need to get at presenting your projects professionally and this might include things like a professional project proposal. It might mean that you need to articulate risks involved in your project, say data privacy or confidentiality concerns and how you’re mitigating those. You might not think about data scientists as project managers but in a sense, you are doing project management in all of your analytics projects. Hence, becoming a better project manager makes you a more effective data scientist.
So, how do you improve your project management skills? It’s simple. Try to be a project manager for events in your university or club events such as hackathons, career fairs, webinars, etc… That will give you experience on how to plan efficiently, delegate tasks to other members, and communicate effectively. You will learn how to change your plan of action and adapt quickly if things don’t go according to your initial plan. This flexibility and ability to handle uncertainty will come in handy in your data science career.
Leveraging Soft Skills for the Future
To sum up, I hope my article has proven the significance of soft skills and how they beautifully complement your hard/technical skills. These are just some of the skills that will surely help your data science performance. Developing soft skills not only provides value to your organization but can also leverage your career exponentially.
It is said that most technical skills can become obsolete in the future due to automation, but one can be sure that soft skills are different. These are hard to be replaced. You can then be confident that such skills will not disappear overnight. So what are you waiting for? Better acquire and leverage them now.
Want to read more content like this? Follow us in our digital accounts and hear updates from the Asia-Pacific Youth Data Society (AYDS):
- Facebook: www.facebook.com/aydsdata
- Twitter: www.twitter.com/aydsdata
- Instagram: www.instagram.com/aydsdata
References:
