Over the past few years Data Science has become very high paying and rewarding career choice for many technically skilled and adventurous minded people. A lot of data scientists have Ph.D’s in stem fields, and an obvious question that comes to mind is does a Ph.D actually prepare you for a career in Data Science.
It does prepare you in some ways !!
The tenacity to solve problems, the ability to handle undefined problems , the resilience to failure are character traits of Ph.D’s that help working in a field which still taking shape and even now there is a lot debate in the skill sets necessary to become a data scientist. Additionally a lot of fields such as Astronomy, Neuroscience, High Energy physics require handling large amounts of data , machine learning, data modeling, parameter estimation and image analysis to perform scientific enquiry. These skills are invaluable to data scientists. Additionally most people use python or R for some of their analysis and some kind of big data tools as well. A Ph.D. also helps develop presentation skills to large audiences which become very useful in real world situation.
And then in many ways PhD. might not prepare you!!
In scientific world going very deep into a problem and getting very accurate results is important. It is possible to spend a decade to build an instrument, gather data and perform analysis to reduce the uncertainty of a parameter by a few percent. The pace in the industry is much faster, especially if you are in a startup or a small company.
Very often “good enough” is better than “perfect” .
As a data scientist it will be your job to make that call.
In industry data is generated through different types of interactions and often machine learning models are build on data from existing systems. Once the power of using a particular type of modeling proves to be successful investments on creating new infrastructure to get more data might happen depending on business goals and the revenue generating power of modeling.
Another very important part of the job is communicating the power of your work effectively. This might mean investing time in understanding audience needs and fine tuning your message for different audiences such as business or engineering. There maybe a cool technique that is exciting, however you need to understand how it improves product or adds to business goals in the long run. As a data scientists you wear many hats, for some projects you maybe required to produce results in a regular basis and for others you might be building systems that become important in the long run. It is important to know which hat you are wearing and clarify if necessary.
A lot of these soft skills develops during the first year of your career, so it is important to find a job in a team that has the bandwidth to hire a newly minted Ph.D. and has enough patience for the data scientists to learn the ropes of the game.
Something that helped me a lot during the transition process was to hang out with Data Scientists who were already working in industry. Additionally, I visited companies to figure out the day to day tasks of Data Scientists and attended meetups, conferences and spoke to business and product people about their process. This helped prepare myself mentally both for the interview and the job experience.