Data Science, Data Analytics, Data Engineer!!! Seeking a career in Data Science.

In the midst of all confusion regarding the optimal path in the field of data science, let’s take a look at the most popular options along with some advice on how to proceed.

Where to begin? Which path is the best way? The question that all Data enthusiasts face.

The path of growth and learning in the field of Analytics can be accomplished by the following mentioned cases:

Courses:

Show me the Way! There are a plethora of options available when it comes to learning data science. Lately, there have been an ever growing collection of MOOC’s that can help individuals expand their knowledge base. However, this makes finding the right fit for you slightly difficult. Essentially, what makes an on-line course stand out is the instructor. It is a good way for beginners to get started.

Learn By Doing (Kaggle):

It is all about the Accuracy! There are variety of problem statements available with the requisite data. Though it instills a competitive spirit and might require some creative effort in terms of feature engineering and model building, one issue I have observed is that analysts often forget the Business value/impact of the problem being solved. Most competitors simply fit a GBM or Random forest or fork someones kernal that gives good accuracy without taking the time to treat it as a proper project and carrying out any steps to pre-process the data or validate the results. Often, I have seen data enthusiasts not even taking the time to generate features, they often make all kinds of aggregates or copy other kernels irrespective of what sense do these features make.

Learn By Doing (From Scratch):

Climbing the Grand Canyon, the image that comes to mind when I think of this approach. This involves everything from finding a problem statement that you are passionate about to carrying out all the steps in the data pipeline. Though the task may look insurmountable however the pay-off makes it worth the hard work. This way you can ascertain much more knowledge about all the facets of a data science project. To start with, approach a Business as a researcher, understand their Business issues and then formulate a methodology on solving the issue.

Choice is Yours!

My preferred method is to start with the right MOOC where the instructor teaches the topic in an interesting way. With each topic, research over and above the MOOC material is necessary to truly grasp the concept/math. Following which implementing what you have understood on a Kaggle playground competition is a good idea. One can learn a lot from other kaggle kernels as long as sufficient time is dedicated to building understanding, why a step was taken. This will cover all aspects of learning the concept.

Keep in mind, that the path of a data science journey takes a lot of study and hard work.

Leave a Reply