There are a lot of skills that will help to become a data scientist. These includes technical skills such as machine learning, deep learning, mathematics, programming and others. But in all these, there are also some non-technical skills or principles that data pros should learn.
Talking to subject matter experts
As a data scientist or one aspiring to become a decent one, you should have the courage to talk to people. Information is king. And in different lines of work, there will be various people who specialise or experts in their own fields. Make it a point to talk to these individuals, they are valuable resource persons.
You will not have access to experts all the time, during these periods you should try and put yourself in the shoes of others. With cognitive empathy, we are not trying to put ourselves directly in their situation but understand the thought process on how certain decisions and data came to be.
If you are a fan of detective literally works like Sherlock Holmes, you would be familiar with this process. Where the detective will try to trace the steps on how crimes happened.
Do not trust the data. Doubt the data. Even if you try to trace how the decision came to be and arrived at the same result, you might want to question it. There is a saying that “Sometimes the data lies to you”. While this is not always the case, in some instance it could just be incomplete. Or you are looking at a different time frame that is outdated. You must nurture a skeptical mindset. A bit of knowledge in statistics will help.
From our partners:
There is a saying that curiousity killed the cat. But remember that the cat has 9 lives. We are not saying you commit 8 mistakes and leave it at that. Try to have a creative thinking when you don’t have access to data. That is to check what variables or missing information could be missing. Ask why the data is like that. There might be a missing attribute that is actually there but was not given to you thinking it might be irrelevant.
Your job is not only the data aspect of the project. Remember that you would need to communicate not just with the members that are part of the data science team, but also to upper management. You would need to be able to explain without using technical jargon. If it is unavoidable to use these terms, explain it further by giving examples or something to relate to.
Another common phrase is “Communication is the key”. They are not wrong, and this applicable to the line of work of being a data scientist. Communicating with the management, team mates and those that are of subject experts are crucial. A misstep that conveys the wrong information to everyone could derail the project in levels you could not imagine.
Having a scheduled discussion will help consolidate ideas and information. Schedule it regularly but not too frequent that it might disrupt everyone. This will provide clarity and understanding on what the status and direction of the project is. Before these meetings, think of possible questions that could be asked, and prepare for answers and possible solutions that you are considering. The most common question is the data integrity. Back these up with data and be honest with everyone.
Thank you for Z/O Digital