If you've arrived here, I suspect you're eager to learn what it takes to become a Data Scientist.
I'm creating this essay because I see a lot of folks in a pickle, and there's no information out there save for many articles on which online courses you should do.
On this particular question — Should I get into data science? — I have received interest from people from almost every background — IT, mechanical, electrical, electronics, energy, chemical, and civil, with people from B.Tech, M.Tech, B.Sc, M.Sc, and even PhD, ranging from no experience to 5 years of experience in my own circle and outside.
Hence, using my understanding of economics, psychology, and study techniques, I will do my best to reveal how you can make it.
Economics
It is my fervent opinion that you should understand how this may work out economically. Others will assess you based on your history when you begin applying. As a result, they will assess your market value and issue an offer letter.
If you are not an IT specialist, you will have to put in a lot of work to learn it, and you will not be disappointed! Hence, unless you have a degree from a Tier -1 college – Bachelors or Masters — you should be prepared to be dissatisfied if you earn more than 7 lakhs*.
If you work in information technology, you are in a good position. We really value candidates who have worked in IT in DS interviews since we don't have to educate them the intricacies of how IT works.
Regardless of how sophisticated DS appears, it is still IT at its core. We want individuals to understand IT concepts such as database operation, querying, ETL, testing, deployment, and clean code. Individuals with a background in IT have had exposure to this, which helps their résumé. The only thing they have to worry about is all of the data science, which appears lot more doable to them than it does to non-IT folks.
Because their learning curve will be shorter and they will be able to utilise their previous skills to receive a higher offer, the risk of exploring this option will be much reduced. Thus, until you make 7 lakhs or more in India, you can do it. Even if you receive less than 7, you will easily catch up in following years.
Psychology
To be honest, entering the field of data science is a game of catching your worries. It will be intimidating at first, with so many acronyms and jargons, but you will have to grow used to it. And if you were afraid of mathematics in college, it's time to overcome your phobias or give up on your goals of being an engineer. My early days were riddled with uncertainties. The learning curve was difficult, and confusion only increased. After a while of taking copious notes and rewriting them three times, I was able to assimilate information on a grand scale.
Study Methods
It's often helpful to have a friend on long excursions like these. I met folks through Facebook, Whatsapp, and Telegram groups and learned a lot from working on projects with them. Work on the same project, contribute code to Github, and talk about it. This will keep you going and allow you to broaden your techniques.
You'd be shocked how many different solutions there are to handle the same problem. A skilled data scientist is someone who has made enough mistakes to understand what will not work. So form groups and collaborate on various projects. Google how others solved a Kaggle challenge and try to grasp it. If you run out of ideas, simply grab the data and accessible code from Kaggle and redo it line by line.
This is how my first clustering project went down. I merely reworked old material and attempted to understand each line and the math behind it.
Then, with the assistance of StackOverflow, I began creating my own. Now, if I'm working on an old problem, I know what to do without any help or lesson. It's a hell of a voyage. Additionally, unless you perform the same thing every day, you will forget the syntax. Therefore don't be concerned. Simply launch the docs or instructional and begin writing.
Integrate yourself into the ecology.
Invest time in your JobCurators profile from the start. It fulfils two functions. Not only will you begin networking with professionals in the business, but you will also learn about DS projects and the most recent breakthroughs in the area.
Because DS is changing so rapidly, you need a reliable source of information, which LinkedIn provides. This Facebook group is also highly active, and you may utilise it to connect with individuals who share your interests.
Over time, you'll realise how little you recall of what you read. As a result, devote time to taking notes. I used to pause Andrew NG videos and take notes. That took about three times as long as viewing tutorials, but I ended up learning more – which is necessary in the beginning.
Even if you are not an expert, try to answer others' questions. This will result in greater clarity on themes. Several of these may also serve as interview questions.
Courses
There are several courses available, as well as numerous approaches to the same themes in various methods. First, I believed that you should only take one course, which meant that you should locate the finest course and take it, and I chose Andrew's. Later, when I was situated and had leisure, I looked into the Udacity course.
I discovered that it is a more practical lesson on the advantages and drawbacks of algos, which was not covered extensively in Andrew's course. Hence it appears that various specialists have different topics to discuss. As a result, if you wish to begin a course, just do it. These are all free to download.
If you are unhappy with the course's approach or material, you should stop taking it.
I agree that Andrew's course is a bit thick and takes many viewings. Yet it is how you learn not to give up and to do what is necessary. If data science was simple, everyone would have done it, demand would have been less than supply, and individuals would not have been paid so much. Therefore, start with whichever course you choose and don't give up quickly.
Time
Many individuals wonder how long it will take to prepare and find work. Because time is determined by your present knowledge and understanding capacity, I like to describe it in terms of tasks.
Completing all of the fundamental courses, as well as 8 supervised and 2 unsupervised projects, can easily take 4-6 months of committed (10 h/day) labour.
You may easily do it in 8-12 months if you work part-time. (This includes the time it takes to discover firms and interview with them.)
Interviews Ideally, after you've covered the fundamentals, you should begin interviewing with firms to gain a sense of the structure of the interview and become comfortable failing at it. On platforms like JobCurators, you may locate some decent firms recruiting.
Note that these interviews might be really difficult. My experience has shown me that the better the firm, the more difficult the interview. The depth of the interview practically functions as a proxy for the team questioning you. Therefore, if you're having a simple interview, odds are you'll end up producing some low-quality excel or scraping. And if you want to secure a decent job, you need be skilled at navigating an interview.
As you start interviewing, some of them will assign you coding homework. It is essential that you develop your own code and review it in order to uncover faults.
I can't tell you how many mistakes I've made in these tasks. I created projects with defects in the code and bad development techniques. Yet with each failure, I learned how to do it better the next time.
As I look back at those scripts, I realise how far I've come. One thing to keep in mind is that it should not be done by your pals. You can consult them if you wish, but make it a practise to solve difficulties on your own. That is easier said than done, yet it will help you develop character.
Everything you should know to improve your chances of getting an interview.
Statistics – Many businesses want to know about the Bayes theorem and the Normal distribution.
Basics of machine/deep learning — Pros/cons of algorithms and how they function
Python + Competitive coding skills are required.
SQL knowledge are necessary as a minimum. It's useful to be familiar with both SQL and NoSQL databases.
Cloud computing is a valuable addition, but it is not a must-have. Discover AWS (Amazon Web Services)
Github – Showing off good work on Github demonstrates confidence and excitement, which is what the top organisations seek for.
Blog – Blogging allows you to gain self-awareness about your hobbies.
Additionally, because I learned a lot by reading other people's blogs, I enjoy sharing my own knowledge.
How I evaluate businesses:
The more difficult the interview, the better the team, work, and pay.
I look at the LinkedIn profiles of team members and leaders. I go through their work history and current job descriptions. Sometimes individuals give imprecise descriptions or say I scrape — this is a good signal that I should avoid the firm.
Data science is enormous. Join a team with brilliant individuals if you want to learn rapidly. They might be startups or multinational corporations. Another blog article may be written about the dispute between startups and MNCs.
There are lessons to be learned from both of them. Startups are agile, but MNCs have resources.
Attempt to work for a firm that is a market leader in at least one area and has a research-oriented attitude. It should not be a corporation that performs data science to save money but does it because it is their bread and butter. Such businesses, however, are uncommon.
At the end of the interview, I inquire about the team's age, size, and average experience. I don't dislike startups or small teams; I simply like to know the numbers. I also inquire about the challenges they are currently working on, but most of them will not respond due to privacy concerns.
I have JobCurators Registered profile, so I can look at the company's hiring growth over the last three months, six months, and a year. This is especially important for small businesses and startups. The health of an organisation is closely tied to the growth of a team. This is a useful but not required condition.
Check out the company's reviews on Glassdoor.com. Check to see if the work environment is safe, and if not, cancel the procedure. If everything appears to be in order, be prepared for the type of CTC that may be implemented. You may also verify the figures ahead of time to see whether they fall within your price range. To save time, you may also ask them directly what their range is.
Indeed, this is all one lengthy narrative. Just give it your all. You must earn it.
Please let me know what you think about the article, and please share your own experiences thus far.
*In India, the median income for a data scientist is 7 lakhs. You should not anticipate more than this unless you are from a Tier 1 college, have more than a year of experience in DS, or have extensive IT experience with excellent DS understanding.