How I Switched to Data Science

How I Switched to Data Science

This is very common to switch to data science. Most data scientists I know out there do not have a degree in data science. They switched from another area. I also know many people who are trying to switch from another major. I meet many people being confused if it is the right career track for them. Well, the decision is yours. Everyone may have a different journey and may have different opinions. In this article, I decided to share my own experience. For some people, it could be an interesting read and for some people, it might be helpful information.

It was a long process. I am starting with my background a little bit. If it feels too long for you, please feel free to skip to the data science area after two sections.

Background

I do not have a data science or computer science background. My bachelor was in Civil Engineering. I only knew Microsoft Office and a little bit of Auto CAD as my computer skills. I worked in different construction companies for five years and never liked it. Then started my master’s in Environmental Engineering in a public school in Texas. I worked as a research assistant for almost three years. That was a nice experience. After that, I was in the teaching profession for a couple of years.

I had no coding experience at all.

Starting of Coding

While I was teaching part-time, I decided to learn to code. Because there is a lot of buzz about how important it is to have coding skills in the twenty-first century!

I started searching for some free resources on Google. Because I had no idea at all how coding looks and feels like. I just wanted some free experience before spending any money on it. I came across the LaunchCode website. They had some free courses and practice platforms. I even didn’t have to install anything. The first course was an HTML course. As I did not have any idea and even did not talk about it to anyone, I started with HTML. When I learned to code a few lines of HTML and I saw some output, it was very thrilling. So I kept doing it. After HTML, I also took the CSS and JavaScript courses for free in the LaunchCode platform. Now I know that all was just introductory level courses.

At that time I was so naive that I thought I had learned enough to get a job and earn a lot of money!

I started applying for jobs and for sure didn’t even get an interview.

One morning I received an email that LaunchCode is offering a free BootCamp in Miami that is called CS50. Miami Dade College will host it. That was Harvard University’s introduction to the computer science course. They videotape their classrooms and LaunchCode uses that for their Bootcamp. That course is also available in edx and can be taken for free using the audit option.

I made a community of coders and got familiar with the essentials of coding like data structures and algorithms. They also teach some HTML and CSS. For the first time, I used SQL there. That experience gave me the realization that I have a long way to go before I can call myself a software developer. Yes, after LaunchCode, that was the goal. I wanted to be a web-based software developer.

I secretly wanted to work in the Data Science and Artificial Intelligence industry but I thought that’s probably too hard. I thought I needed more coding expertise before I can start that.

After Bootcamp, I kept practicing PHP, SQL, JavaScript, and WordPress! I was still teaching and was hating my job. I was desperate to find another career. I think most people start thinking of a different career because they are not happy in their current job or not making enough money. And then because of a lot of buzz about Data Science, focus shifts there.

I took more classes in Udemy and Udacity and was applying for jobs as a web developer. I worked as a web developer in a startup for a couple of years. While working there, I thought I probably know enough to try Data Science now.

Beginning of Data Science

I started learning Python in Udacity. Then I took a series of courses in Coursera:

Applied Data Science with Python

Natural Language Processing in TensorFlow

Convolutional Neural Networks in Tensorflow

All these courses teach how to use different libraries to analyze data and make predictive models. But I wanted to learn to develop the machine learning models from scratch, not only from a library. I found another great machine learning course offered by Professor Andrew Ng of Stanford University.

While taking this machine learning course, I realized I need to develop more statistics knowledge. I took a probability and statistics course in college. But that wasn’t enough. This course on Coursera helped me a lot:

Statistics in Python, offered by the University of Michigan.

It has three courses. The first course is more about exploratory data analysis. The second one is inferential statistics and the third one is model fitting.

As you can see I use Coursera a lot! It has a lot of great courses. But it takes some time to find a good course that is suitable for you. A lot of time I started a course and after I have done it halfway I realized this is not for me. So, there will be some time laps there. If you are totally new to Coursera and do not know how to take courses for free there, here is a six minutes tutorial for that:

After taking some of the courses, I did an internship and a few projects. I attended some conferences, presented some personal projects there, joined several Meetups, and made some good friends, and learned a lot from them. Currently, I am doing my MS in Applied Data Analytics at Boston University.

Image for post

Image by Author

If you are a beginner, you may think, after learning that much why I felt the need for doing a master’s.

I will write a separate article on that. But if you ask me, if I find it useful. My answer is yes. I am loving it. It was a great decision. On the other hand, If you ask me if it is possible to become a successful data scientist without a degree. My answer to that is also yes. I have met many data scientists in the meetups and data science conferences who do not have any data science or a computer science degree.

The Mistakes I made

The first mistake I made was getting scared. I spent so much time learning web development and also worked as a front-end web developer knowing that I actually wanted to be a Data Scientist. I was desperate to change my teaching job as fast as I could.

Instead of getting desperate and being scared of trying something new, if I would just get into it, it would save me a few years.

The second mistake was to think that data science does not require much programming problem solving or algorithm skills. In fact, some experienced software engineers told me that. But as I went deeper into it, I realized I needed more algorithm skills. I went back and had to learn some more. Boston University’s Applied Data Analytics master’s program starts with an advanced algorithm class. Also, the professor told us, we did not have enough time to go further and we needed to learn more in the course. Now, I am learning more by myself.

These were the two major mistakes.

My Suggestion to the Self-Learners and Beginners

I do not claim to be a hugely successful Data Scientist yet. But after I spent that much time learning, I think I am eligible to share some suggestions.

  1. There are a lot of boot camps out there, charges 10000 USD for their courses and just teach you the very basics. But they will sell you the courses by saying that they will make you a data scientist in 12 weeks or 4 months. Do not fall for those traps and spend your money there. If you can find a Bootcamp that tells you they will teach you programming and algorithms in 4 months that is the best starting point.
  2. If you really want to be a data scientist and grow in this profession, you will need a lot of different skills. SQL and NoSQL databases, statistics is essential and programming as I mentioned earlier. These are the basics. Spend enough time to learn them well. Rushing will never help. Here is an article on that.

3. If you do not have any math or technical background, you may have to spend some extra time. But there is nothing to be pessimistic about. If you want something, you need to spend time and put some effort into it.

Data Science is a multidisciplined zone. This profession can use people from different backgrounds.

I know people from a business background doing very well as a data scientist.

4. Network, network, and network! Join different data scientist’s meetups, seminars, and conferences. It will give you a lot of perspectives and ideas. Also motivations. Only thing is, different people will keep giving you different suggestions. Don’t lose your focus. You need to make a track for yourself and stay on that. Because everyone is different. Whatever worked for somebody else may not work for you. So don’t change track seeing someone else’s experience. That way you will keep changing track every other month. Learn a little bit of everything but not enough of anything. That happens to a lot of self-learners.

Conclusion

I shared my own journey, some great resources I used on the way, and some of my ideas in this article. I hope it will be helpful for some of you. I am sure many people will relate to some of my experiences. Please feel free to ask if you have any questions for me.

Feel free to follow me on Twitter and like my Facebook page.

 

#DataScience #MachineLearning #ArtificialIntelligence

Leave a Reply

Close Menu