Thinking about transition from electrical engineering to Data Science?

machine Learning/Data Science become so popular and demanding in the job market, a lot of people start to think about transition to this new field. Some of them have no solid statistical or data science background, they may come from different fields, electrical engineering, mechanical engineering, aerospace engineering…, If you are one of them, this article might give you some insights about how to prepare for it. 

Do you really know what is Data Science? 

Before jumping into any my learning experience and advice, please do research this most important thing first! Here are the step by step instructions: 

Step 1: Copy and paste it in Google and read at least 10 articles about this answer, either they want to promote their bootcamp course or just sincere answers from current data scientists (You can search in Quora). 

Step 2: Search through Linkedin about data scientist in different companies, and add them as friends. Read their profile and learn what are their responsibilities. 

Step 3: Talk to at least 3 data scientists in your friend circle about their daily job and ask them what are the most interesting and frustrating part(if any, they might in love their job :D ) in the work. 

Step 4: Summarize all above and list down all pros and cons for YOURSELF. 

Once you complete above and still have passion to jump into this field, please go for it and continue reading! 

Fundamentals

One of the most important thing for all jobs and fields is fundamental knowledge. Please note that this is not only about the theoretical knowledge, but also the hands-on experience that at least you have heard about. For a general data scientist, here are some of the most important fundamentals. 

1. Statistics and Probability 
2. Machine Learning/Modeling 
3. Programming Language 
4. Product Sense 
5. Presentation Skills 

I can talk more details about each item, but the key take away is you should know(NOTE, not to ask you specialize) all of them before reaching out to any recruiters/referrals! 

You can study item 1 and 2 through so many online resources, i.e., Coursera, Udacity, Stanford, and even Youtube (A lot of great classes have free videos uploaded), and one of the most popular one is Andres Ng’s Machine Learning course, which gives you a very great overview of machine learning techniques, if you want to study more deeply, you can also consider to take a course in stanford, the actual one is much more intense than the Coursera one. 

As for the programming language, the most frequent question for a beginner is:What should I know first? SQL, R or Python? 

This is very case by case and depending on different job and company. As a general rule, you should be proficient in either R or Python and optional on SQL to be a data scientist. Again, it depends on what is your next job role. 

Item 4 and 5 are kind of hard to have a very distinct change in a short time, but you should always practice it as soon as possible since this might be the distinguished aspects if two individuals have similar technical background. 

Lastly, I highly recommend to register one bootcamp if you don’t know anything about the above, they can walk you through and you would also study the fundamentals there in the same time. The most famous one is the Insight Data Science, but this is pretty hard to get in as far as i heard. And they prefer solid background PhD, so it is not that accessible for a lot of transferred people. 

Go deep and behave as an interviewer/interviewee 

After learning your fundamentals, you should really try to hone your skills. Start from very simple questions to hard and tricky questions, just behave like interviewer and interviewee at the same time. The most beneficial thing for me is I wrote down an 18 pages of interview questions and answers by myself. I try to summarize all technical and modeling questions, and wrote down my answers. Through the process of finding the best answers, you would go much deeper for every one! Be sure to practice all of them and truly understand! 

Build your Data Science experience 

Once you have a very good fundamentals and understanding of each method, you should start build your data science experience as soon as possible! 

Start search data, download data, clean data, feature engineering, build model, model evaluation(and go back to cleaning/feature engineering again). Once you have at least 4-5 times of this end to end process, you would know much better about what is data science and what you actually can do in your work. 

Remember to think always like an interviewer when you find any solutions, in the end, data science is about the ability to find insights from unreconstructed/uncleaned data. So you should always ask why instead of just doing it without thinking and comparing. 

Last but not the least, start to interview before you feel you are ready, I did it within three months and I was not even 20% confident, but you should always push yourself from your comfort zone and learn lessons as soon as possible 

Good luck and have fun in Data Science :)