When a field of expertise or a discipline becomes too popular, a halo of expectations are built around it. Data science has a similar one. The role of a data scientist is marketed as one of the best jobs the world has ever seen. Of course, it is a very opportune field with a lot of interesting stuff to do and learn. But there are always two sides of the coin. Today, we will try to see the flip side of the data science coin and understand what it is really like to be a data scientist. Thanks to a bunch of data science professionals with different degrees of expertise and experience who have shared their thoughts on the matter. So, before you start your data science training Let us see what it looks like from the inside.
A young data scientist who has just spent a year in the industry says that he has had to read research papers, write algorithms, then code those algorithms up, on the same day. Forget the degree, just the variety of this sort of skill set is frightening. On the brighter side, this person loves his job and would not have anyone else to do it for him.
A firm grip of mathematics and statistics is necessary if you are going to shine in data science. But at the same time you will need coding skills. You will need a basic understanding of the principles of machine learning. And if you are working on a fully fledged artificial intelligence system you may need more than just a basic understanding.
So, it appears that it is really important to keep yourself open to a varied platter of knowledge.
According to the experience of most data science professionals, you never really graduate. That is to say that you never stop studying if you are working as a data science professional. Reading research papers to improve your game is like a part of the job. Analyzing a business problem and then finding the best approach towards a solution; arguing over it with the co-workers, it all has a structure like that of a college group that submits assignments together.
It would have really been like college had not there been a business involved. Businesses hire data scientists with the expectations of results. Results like knowing their consumer base better; Making their marketing spearheads sharper; increasing efficiency of the supply chain. Now, all the data scientist has to work with is data and a very limited amount of time. Data scientists have to look for easily applicable and scalable solutions to the problems rather than looking for the perfect solutions which may take more time and effort to implement. So, yes, it is business before science.
A large part of the data science professional’s time is spent in preparing the data to feed into algorithms and before that to train the algorithms. This part is so important that they cannot really leave it up to someone else. Popular notion is that data scientists like to clean their own data. The quality of the data determines the quality of results as well as the reputation of the data scientist. Data scientists would give anything to code all day or to build models all day, but they just cannot skip this part.
A data scientist’s job does not end at creating the models or finding some results. It starts there. They need to bring the findings to the management and convince them to act upon it. The data visualization is a crucial part of a job, and it has to be done well. Also, they often have to work with software engineers to create better code to realize their visions. Being able to express that vision makes a lot of difference.
As a data scientist you can do a lot of cool stuff but the business requirements do not let you delve too deep into anything which is both a good thing and a bad thing depending on who you are. Your success is determined not by how good you are at maths or statistics but by how well you manage your skills and organize your job. Al and machine learning are very important subsets of data science but they cannot replace it.