3 Pieces of Advice for Aspiring Data Scientists

Original Source Here

Data science is like the shiny new object in town that people are easily attracted to. But soon after you are into the journey, it can feel overwhelming, exhausting and sometimes you feel clueless, not knowing which direction to go. The more you learn algorithms, programming fundamentals, and the libraries, the less confident you get in your ability to push through. Imposter syndrome can be chronic in data science.

The following are my 3 pieces of advice to fight back.

1. Find your niche

Data science is not a single course on a single topic, it’s truly a multidisciplinary field. The list of things you come across can be overwhelming, and they are all over the place across many different areas. Some examples:

Probability and statistics: Central tendency, distribution of data, hypothesis testing…..

Programming fundamentals: Python, SQL…..

Basic data science libraries: Numpy, Pandas, Seaborn, Matplolib…..

Machine learning fundamentals: Feature engineering, bias-variance tradeoffs, cross-validation….

Regression algos: Linear, LASSO, Ridge, SVR….

Classification algorithms: Logistic regression, random forests, boosting techniques….

Clustering algorithms: K-means, hierarchical, DBSCAN….

Deep learning: CNN, RNN, Tensorflow, Keras, PyTorch….

Forecasting techniques: ARIMA, VAR, LSTM….

Engineering: Product deployment, drift monitoring, retraining algorithms….

Field of application: Computer vision, NLP, decision science….

Subject matter specialization: Customer analytics, FinTech, recommender systems…

The list goes on and on!

The question is — how is it possible that one person becomes an expert in all these areas in merely a year or two, through self-learning or a Bootcamp? It’s not, and that is why imposter is real in data science. The more you learn, the more you think you don’t know anything; you never feel ready to move on to the new chapter with a new job.

The reality is you’ll never know it all, it’s an illusion.

So here’s my advice. Find your niche. And start to specialize in the things that you are interested in. Choose which field of application excites you most; which algorithms you are most comfortable with? Do you want to be in the field of NLP? Or are you interested in decision science?

By specialization, I do not mean that you should abandon everything else and focus on just one thing. Rather I am saying that you should spend ~80% of your time on the area you are specializing in. And keep the remaining 20% door open for learning the other areas that you also find fascinating.

2. Teach what you’ve learned

Teaching means learning twice — literally. You learn the first time while doing research for teaching and the second time when you are actually teaching.

By teaching I do not mean the traditional classroom style teaching; teaching can take many different forms these days. Arguably the easiest is writing a blog post or an article. If you write an article on a topic you’ve just learned — an algorithm, a programming concept, or a visualization technique — naturally you have to do further research on the topic. By way of explaining the topic to others, you will refresh your own understanding of the topic.

There are many writing platforms these days, and you can start as soon as you’ve finished reading this article. If you feel like you are not ready to publish on a big platform, you can always start with a smaller outlet or even a personal blog. But I strongly suggest that you let the world see whatever you’re writing, so publish your first blog on any public platform. It forces you to get the best out of yourself and as a reward get immediate feedback through comments, claps, and readership.

But writing does not come naturally to many people.

Some people are just good at explaining things verbally. That’s fine too. Create a YouTube channel and share your knowledge with the world. Don’t care about viewership in the beginning because that’s not what you are after; your primary objective is to brush up on your own learning through teaching. Consider viewership just as a byproduct.

3. Find your network

And the final piece of advice — find a community.

Traditional schooling has many advantages. You have your teachers, classmates, peers, and friends with who you interact every day — we take all these valuable interactions for granted. You occasionally also go out of town to attend conferences, workshops, and seminars — big networking events to meet unknown people and make new friends. In fact, there’s a whole system and infrastructure built to support you.

But if you are self-teaching, learning out of a Bootcamp, or taking a MOOC, you feel like a lone soul, left alone in the desert. There’s no one with whom to share your ideas, feelings, and exhaustion.

All is to say — find your community. Maybe a local meetup group where you can find people with similar interests. Hang out with them in person. If unable due to COVID restrictions, find those communities online, join the weekly/monthly meetups and keep in touch through social media platforms.

Parting thoughts

Data science is an exciting field. But it’s also overwhelming. Things can change real fast. Tools you are learning today may no longer be useful tomorrow. So take data science as a journey rather than a destination. Prepare yourself to be a lifelong learner.

While learning, share your knowledge with the world on whichever media you are comfortable with.

And finally, find your crowd, have fun and help each other out!


Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: