How to be a more productive Data Scientist? Don’t do more, do it smarter!

A few years ago, I put a lot of work into a hobby project of mine: a self-learning Python chatbot that learns like a newborn baby. It was an exciting concept: I wasn’t allowed to pre-define different grammar rules, so the bot had to identify and learn every rule of the English language by itself. A fun project to work on, but it was also very, very complex. I came up against a lot of technical, logical and conceptual challenges.

I remember, one of my biggest obstacles at the beginning was that I couldn’t come up with a good enough structure to store the new words. I was writing and testing different solutions in Jupyter Notebook for hours, but none of them fulfilled all of the requirements I had… then I said, enough is enough. I switched off the computer and went out for a jog. And an hour later while I was having my after-run shower, magically, the solution just hit me!

I like to recall this story from time to time, because it helps me to remember one of the most important rules of being a more productive data scientist: you don’t have to work more, you have to work smarter!

In this article, I’ll give you six practical tips. Regardless of whether you’re a junior or senior data scientist or you’ve just started to learn, these will help you get the most out of yourself and be a better-performing data professional!

Rule #1: Don’t be busy to impress your boss!

My colleagues always made fun of me in the office, when I was staring at my screen or at my notepad for 10 or 15 minutes without hitting any keys on my keyboard. Well, in a busy startup office, pondering seems like daydreaming. But the truth is that data science is a job that needs deep thinking. Without spending sufficient amount of time to fully understand your data, your analysis, your predictive model, etc. – data science often does more harm than good.

A good manager knows that. So even if it feels weird to just sit and think about things, don’t be afraid to do so. Your promotion won’t depend on how busy you look during the day but the quality of the results that you deliver at the end of the day.

So rule #1 is: Don’t be busy to impress your boss! Take your time thinking.

Rule #2: Be fluent in SQL and Python!

If a nicely delivered data analysis is the cake, then SQL, Python and pandas are the mixer, the tray and the oven. As a good chef you have to know exactly how these tools work. I’m not saying that you have to have every function of every Python package on the top of your head (that’s not even possible). I’m just saying that you should build on solid fundamentals.
Writing complex scripts efficiently needs a flow state. If you have to keep looking up the syntax of an SQL JOIN, a Python for loop or a pandas data import, that will constantly break this flow. You won’t just slow down, you will lose motivation.

So if you don’t feel 100% confident with SQL, Python or pandas yet, take a few more tutorials and more importantly: practice, practice and practice – a lot!

practice SQL for data analysis

Rule #3: Mute your devices!

I like companies where working remotely is supported. My best practice is to work four days on site, but take one day a week to work from home (or from anywhere else). For me, this worked really well. While being at the office to meet and to communicate with co-workers is really important, I always felt that my most productive days were when I worked from home.

Because what does a data scientist do? She comes up with statistical models, writes code, tries to understand the business impact of the analyses… These are very introverted tasks. To have full focus, you have to be alone in an interruption-free environment.

The key is interruption-free. So working from home won’t be productive, if:

  • your colleagues ping you constantly on Slack, in email, on Skype, etc.
  • your boss is checking on you every 10 minutes
  • your friends send you funny videos on Facebook, etc.
  • you schedule the housework (laundry, grocery, repairman, etc.) for your working-from-home days

When you work, you work.

It’s better for everyone if:

  • by default, you mute all your devices. You can take 10 minutes every two hours (let’s say) to follow up with ad-hoc requests by your colleagues or your boss.
  • you have a room where you can sit alone and even with family (if they are home) you clarify that now you have to focus.
  • you do the laundry before or after work and not during the day when you feel lazy about your data project.

As you can see, the emphasis is not on working from home. It’s on working uninterrupted. So if your workplace is more strict about remote working, you can still book a small meeting room for yourself at your office or you can go to a coffee house — just make sure that you find a place where you can focus.

Rule #4: Sleep enough!

A few times a year there is a big project, a big campaign, a big release – at every company. During my junior years, I was really stoked by these periods and very often I took on 60-80 hour work weeks. Being motivated is a great feeling. But there are still only 24 hours in a day. If you don’t have time, the easiest way to spare some is on sleep time… But at a point, I had to realize that not sleeping is the worst strategy.

Your brain is not a machine. If you don’t sleep, it will underperform. It doesn’t matter how much you work, if you do it slower and make more mistakes. It’s just not effective.

Actually, even machines have to be restarted, otherwise they slow down. You have to empty the RAM, shut down background processes and get back to start state, otherwise the performance will drop. Do the same for yourself!

The Junior Data Scientist's First Month

A 100% practical online course. A 6-week simulation of being a junior data scientist at a true-to-life startup.

“Solving real problems, getting real experience – just like in a real data science job.”

Rule #5: Read books – not only about Data Science

Are you looking for a data science mentor? What if I said that you already have dozens of mentors at home on your bookshelf? Okay, maybe that’s an overstatement, but still: reading a smart person’s well-organized thoughts for hours is one of the best ways of learning.

But books are not just about knowledge. They are about inspiration, too! Thus I always recommend to my course participants that besides the data science books they should read regularly on other topics, too.

For instance, the idea of this very article was born when I read a book called Presentation Zen (by Garr Reynolds). I read something in it that resonated with me and that I could translate to data science – even if it came from a very different field. And boom, the idea of this article was born.

Rule #6: Switch off, have a life, have a hobby!

I know from my own experience how easy it is to be caught up in work… But remember the shower-thoughts story from the introduction!

It might sound surprising, but data science needs creativity. To solve a tricky challenge in your code, to set up an unusual statistical model or to interpret a unforeseen finding: you have to use the creative part of your brain. Of course, I’m not talking about the creativity that contemporary artists use to paint black triangles with red circles on a canvas… I’m talking about problem-solving creativity!

Mihaly Csikszentmihalyi says that every creative process has five phases.
In a nutshell:

  1. Preparation: when you study the problem, you think about the problem, and you experiment with different solutions, etc.
  2. Incubation: when you break your process and go do something else
  3. The AHA-moment: the sparkle, when everything magically fits together
  4. Evaluation: when you evaluate whether your magical idea is magical indeed or just a false alarm
  5. Elaboration: when you implement the idea

Note: If you want to read more about this model, I recommend this book: Creativity.

This is exactly what happened to me when I tried to solve that Python chatbot problem. I spent hours in preparation phase but to let the AHA-moment happen, I needed to take a break.

My point is that incubation is an inevitable part of your working process. Sometimes, it’s best for your work if you don’t work but do something else! So don’t feel bad about going out sometimes, going to the gym, watching a movie, etc… It will actually boost your creativity and your results, too. Just make sure you don’t go from one extreme to the other – the preparation and elaboration phases are important, too! 😉

Conclusion

These are my six tips for being a more productive data professional. (And if you think about it, these rules are easily applicable to many other professions…)

If I had to summarize it in one sentence: always try to find the balance! When you work, work focused. But make sure that you have free time and that the free time is free indeed.

Cheers,
Tomi Mester

Cheers,
Tomi Mester

The Junior Data Scientist's First Month
A 100% practical online course. A 6-week simulation of being a junior data scientist at a true-to-life startup.