If there was a list of frequently asked questions about a career in data science, this very generic question would definitely top the list. This is quite understandable because it is a lucrative career path. You can find a new article on how the demand for data science professionals is unmet in many of the cities around the world on a weekly basis. Data scientist is arguably the best technical job at the moment. The salaries are high; the nobility attached to it is phenomenal. No wonder you want to get into data science. But strangely the jobs are difficult to get. You can see that there is a vacancy and you are eligible on paper for the job and yet you do not end up bagging it. This is a crazy paradox in the world of data science. The big data training institutes will train you for the jobs but at the end of the day it all depends on how well your cognitive abilities, matched with some crazy intuitions and analytical thinking, support you on the real turf where the stakes are high and failure is costly.
Can you solve problems?
Industrial data science is all about it. You identify the problem; try to find a solution that has best chances of assuring better business. Then you look for more problems, problems that cannot be seen, but if solved can enhance business. It is really like running an R&D for a business. The domain in this sense becomes very important – the better you understand your domain the easier it will be for you to identify the right problem at the right time. You do it all basically by looking at all the data you can get your hands on. Here again you need to understand what section of the information influx should be intercepted for which problem.
There are a myriad of tools to handle big data and to write code to manipulate data. You can get new skills at a big data training institute. But remember, “Data science is a practice, not a particular skill set”.
The multidisciplinary armoury of skills
When asked to solve a problem or just to identify the problem you may have to go through a loopy procedure. You may have to
- Write code to collect and clean the data.
- Run statistical analysis to understand whether it can be used to solve the problem at hand.
- Build predictive analysis models with the help of the data.
- Visualize your findings in an attractive and effective way.
The solution may or may not be accepted by the business and you may have to get back to square one; but after having spent so much time on problem solving you would not mind it either. The point to be made here is that mastering a set of tools does not make you a data scientist.
No need of learning every tool in the world
You need to gear up with all you can assemble to find best results. Well, the veterans of the field, or at least some of them, will tell you that they have worked with a handful of tools for a long time and successfully too. While others will tell you that the only way to become a successful data science professional is to learn every tool there is. While the first one would be to rash a suggestion to give to an enthusiast, simply because they might need certain skills on their CVs in order to land a decent job, the second opinion is quite impractical.
The idea of being a multidisciplinary task master does not mean you should be able to operate all the tools. It rather suggests that you should be aware of how each step of the data analysis is taking place and how to optimize the functionalities to enhance productivity within the domain. You should have a conviction of the goal and then you can use your resources in the most tactful possible way to get closer to that goal.
The truth about entry level data science jobs
As a beginner it is little likely that you would be offered the role of a data scientist. Even if you are it would probably just be a heavy designation. Experience is very important in this respect. Even after you come out with a doctoral degree in a quantitative stream of science, it is difficult to start off as a data scientist. Spend some years working in an analytics team. Build your data science projects on real social problems. Get connected to a network of data science professionals, discuss your findings; build your way. It is a long way up.