Soft Skills in Data Science: Let’s Talk About That
There is more to being a Data Scientist than just technical skills. Lets take a look at what soft skills can help you on your way to becoming a great Data Scientist
We all see the job boards and we all see the list of doom when it comes to the required technical skillset. I have seen jobs requiring 15 years’ experience in Spark (it was only released to the public in 2012). But even the reasonable ones have enormous expectations, so tackling these skills can be daunting. Many people, both aspiring data scientists and experienced ones spend their time learning complex modelling techniques… not that they aren’t important, however, it’s worth noting, they aren’t everything.
If not technical requirements, what else?
So, there is a mystical beast that plays a key role in becoming a great data scientist, no… not deep learning, soft skills! Soft skills are what separates the average data scientist from the unicorn. Those able to understand and help stakeholders define and most importantly execute clear, meaningful projects.
Having even some of these skills nailed down will inevitably help you on your journey to becoming a standout candidate at any organisation. It is worth noting, this article contains my own findings and experiences with people I look up to in the data science world.
Understanding
Some of the best data scientists I know might not be the most technical or have the best degree from the most prestigious university, but they are able to apply their knowledge to business problems both efficiently and effectively.
So, how do they do it? It’s quite simple really, they take the time to listen, listen again, and finally, listen some more. They take in the information from their stakeholder or customer, then ask the critical question… why? Generally, stakeholders have an idea of what they want in their head, whether that be a dashboard with KPI’s, a prediction of the next quarters sales or something else. It might simple or easy, but if we just build what someone asks, we are not going to be producing our best work or more importantly, results which have a lasting impact. We need to critically understand the problem space and apply our technical experience to solve said problem in the best possible way. We aren’t going to know what the best solution is going to be, without listening and asking the tough questions.
Sometimes stakeholders can find it off putting to be asked the “why” question, but it’s the easiest way to understand from an emotional perspective how the problem affects their work and how a potential solution could ease a burden or provide value to an end user. “Why” allows us to dive deep into the problem space and define a solution alongside our stakeholder which might meet their end goal in a more effective manner.
Ok, so we know what the problem is, and the real reason; why it is a problem, but how do we think of a solution?
Product/Industry Understanding
“Knowledge is not understanding”: Having knowledge of the product space is a great first step, knowing that a product has specific features and knowing that the customer does something in a particular way is another step. But, truly understanding why a product works as it does, and understanding why a customer has a specific problem is key.
A data scientist should be striving to understand their product and customer as fully as possible, because if they don’t, it will be like swimming up a waterfall. You aren’t a salmon!
Once you have a rigid way of thinking in your head, sometimes you cannot change it, even if you want to.
You have to force a new way of thinking; this is usually by stepping out of your comfort zone and taking the new method head on!
Industry understanding isn’t something you can change overnight, but you can change your habits to enable faster growth in this area. Working with your stakeholders, stepping into their shoes, speaking to customers.
Business Business Business
Yes, we don’t all work in a typical business environment, however if you do, proper business acumen goes a long way to securing trust and support from your stakeholders.
I have worked in businesses of all shapes and sizes, from small scale start-ups to global powerhouses, but all have non-data people who you need to hand hold through the process. It’s not a negative thing, but it is bringing someone along a journey that they otherwise would have ignored.
How do you bring someone on this journey? Storytelling! There are many methods for creating a narrative with your work and a good method I have used in the past is pulled from interview techniques.
The STAR technique is used by interviewees to explain a project or narrative, from the problem through to solution.
Situation
Set the scene and give the necessary details of your project at a top level.
Task
Describe what your solution covers in relation to the problem.
Action
How was the problem tackled, are there any interesting revelations not related to results.
Result
Share what outcomes your actions achieved.
The Secret R: Retrospective
Being able to define what went well and what went badly during a project is key, especially if you don’t want to make the same mistake in the future!
This method is a very simple and basic approach to defining a narrative, however, there are many superior ones.
This article by Ling Wong is a great detailed discussion about how to create narrative and clarity with your storytelling:
Finally if you want more information on the STAR method, check this out. Yes it’s from an interview perspective, but you get the drift.
Being a Team Player
Being a team player means being adaptive to the situations around you. Not everyone might have the vision or scope of your knowledge, but it’s worth noting. You don’t have theirs!
Listen to your team, learn from them, share your knowledge. A great way to do this is to share your work regularly, even if you aren’t working on the same problem. Doing this will help to identify holes in your reasoning, they aren’t there to put you down, they are there to rise you up and help you cement trust with your stakeholders. Doing the same thing with your team will help to create a good team bond and propel the team forward when projects might be a little stagnant.
Secondly, being adaptive to new situations is a key part of this process. Adaptability will enable you and your team to pivot to the most important aspects of projects while keeping the overall objective and scope in mind. It’s important to not just use the same approach to a solution, because it worked last time. There may be new or simpler approach which process a much-improved result. Open your mind to new ideas and adapt to changes in requirements fast and efficiently. Remember stakeholders aren’t always data people and they are not aware of how small changes to scope can have huge impact from a data science perspective. It’s best to approach these changes pragmatically and put yourself in the shoes of the stakeholder.
Curiosity
Finally, be curious! Ask questions, seek out new areas or data blind spots. Sit alongside your stakeholders to truly understand the business landscape and most of all, learn! Not just technical skills but learn how and why customers and stakeholders do what they do.
I have no special talents. I am only passionately curious. (Albert Einstein)
If you use some of the approaches listed above, you will become a much-improved data scientist. Yes, you have spent years honing your craft when it comes to programming, statistics or machine learning, but spend some time learning your customer.
Empathy
I know, I know, I said that the last section was the final part, but empathy plays a huge factor being a great data scientist. Many people take this part for granted, many people are great at it, many are not. Put yourself into your stakeholders’ shoes to truly understand how they solve problems, tackle obstacles and overcome deadlines. It will help you as a data scientist, to keep motivated, as you will understand how your new solution will really benefit and improve a stakeholders work on a daily basis.
Many people have some of these skills, but remember, you are a team, work with others to solve the difficult problems. Develop solutions together as in a team you will be able to create some incredible value for those around you.
I am still working on my career, improving my craft, and understanding what it means to be a good data scientist (I’m not sure if I’m there just yet!). I have a long way to go and I’m sure it will take time, but one step at a time and you can become a great data scientist.
If there is one take away… try to be open minded to the people you are trying to help.
