Biggest mistake whilst learning data science


Userlevel 3
Badge

I’d like to hear everyone’s biggest mistake, regret or oversight as they were learning data science. Hopefully we can learn from eachother’s experiences and help people currently starting to learn data science!


10 replies

Userlevel 3
Badge +1

For me, it was waiting so long to do it! I first considered doing a masters in statistics 2014, but I decided against it due to all the boring reasons like finances, time etc. etc. I came back to the idea in 2019 and I did a masters in data science part-time, alongside my job. Being able to do it part time made going back to university way less daunting and although it was tough, I absolutely loved all the learning.

I’m not saying a masters is necessary to do to get into data science, but for me, I think it was. I needed the structure and discipline of the masters to get me started on my career in data science. 

Userlevel 6
Badge +2

My biggest mistake was avoiding statistics so much while I was studying! I did maths for my undergrad and we did statistics courses in 1st and 2nd year and I decided that statistics was of no interest to me so just stuck to pure and applied maths for 3rd and 4th year. I loved pure maths in particular so don’t regret that so much but I really could have done with more of a statistics background when I started studying data science and later working in it. 

Userlevel 4
Badge +1

If I could give my early-DS-self some advice, I’d say get going on something real sooner!! I spent a long time on codecademy learning everything I could about python, but I think you always learn more by getting stuck in and making some real mistakes.

 

Google will always be there to answer (most) coding problems, so I didn’t need to be as worried about not understanding literally everything. 

Userlevel 3
Badge +1

As someone that got into Data Science in a “non-traditional” way, late in career via an Analysts career, I think a lack of appreciation for understanding the underlying concepts for how things work and being able to explain why. It was only as I moved through my career this became particularly important and meant I had to do a lot of returning to previous notes to get properly up to speed. 

I’m not saying have this to hand for each and every algorithm and technique, far from it, but the general framework (plus details easy to get at) is key.

Userlevel 3
Badge

My biggest mistake was avoiding statistics so much while I was studying! I did maths for my undergrad and we did statistics courses in 1st and 2nd year and I decided that statistics was of no interest to me so just stuck to pure and applied maths for 3rd and 4th year. I loved pure maths in particular so don’t regret that so much but I really could have done with more of a statistics background when I started studying data science and later working in it. 

I had a very similar experience! I only chose pure modules during my undergrad, because those were the ones I enjoyed the most 🙈. Wish I knew back then that I’d want to do data science, but oh well...

Userlevel 4
Badge +1

I agree with Simona and Sorcha (in different degrees), I should have carried on with doing Econometrics in my undergrad. I always really enjoyed it, but I had a underwhelming lecturer in second year so I didn’t do the third year module.
It would have improved my understanding of forecasting/regression, and overall increased my statistical fluency. 

Userlevel 3
Badge +1

Underestimating how important good data cleaning and exploratory data analysis skills are and trying to jump straight into complex algorithms. It doesn’t help that in most data courses you learn on perfect data sets that just don’t exist in real life! 

Userlevel 4
Badge +1

I think my biggest mistake was not to get stuck in to a “real” data science project sooner. Learning the theory behind all sorts of fancy algorithms and being able to play around with perfectly fine example data is one thing, but it meant that I felt awfully ill-prepared when I sat down in front of the first unfiltered csv file and initially had no idea how to apply the tools that I had learned. 

Userlevel 3
Badge +2

My biggest problem is something I still struggle with now: not writing things down. When I have a problem, and it takes ages to get a solution, writing that solution down and how you got there helps for next time. I just assume that I’ll remember it perfectly in future, then inevitably I don’t!

Userlevel 4
Badge +2

Perhaps a controversial opinion, but I can’t learn from reading textbooks…

I made the mistake of blindly following my universities recommendations of reading textbooks to learn Data Science, but I always struggled to absorb the content. It took me too long to understand that my optimum learning approach comes from reading relevant texts and applying the content to a project I’m working on. This may mean I don’t progress through books as quick, but the value for me comes from exploring around a topic before moving on

Reply