What Sometimes Goes Wrong When Learning Data Science

career advice job hunting portfolios Jan 10, 2023
Robots who are just about to finish a race.

Data science is an ever-growing field that requires dedication and hard work to master. Sometimes learning data science can feel like an impossible race. With the right mindset and plans you can win that race!

With the right mindset and strategy, anyone can learn data science - but it's easy to make mistakes if you don't know what you're doing. This blog will discuss common missteps on the data science learning journey and provide tips on how to avoid them. We'll explore big-picture topics such as how learning data science cannot be a linear process and how not to use GitHub as your only portfolio platform.

Before concluding we'll also discuss a few more pragmatic topics such as setting realistic goals, staying organized, and asking for help when needed. So whether you're just starting out or already have some experience in data science, this blog can help guide your journey towards becoming a successful practitioner!

Learning Data Science Is Not Linear 

The first major common misstep in learning data science is: Trying to learn like you are still in school. You cannot learn data science like you are in school. Most schools teach most subjects in a linear fashion. Try this analogy for example, learning to speak Spanish could you say:

I'm on a mission to discover the language, one letter at a time. From A-words today to B-words tomorrow and eventually Z words in just a few weeks - I'm confident that soon enough I'll be fluent!

Wrong! Anyone who has learned a new language knows this is not how it works. To become fluent in a language, you need to practice speaking it and make plenty of mistakes, but especially in our modern world you don't need to worry so much about those mistakes—you can always ask for assistance from Google Translate or more experienced speakers.

You often learn data science by starting with statistics (maybe correlation or regression analysis). Then you find a problem and apply the statistical knowledge to that problem. You make mistakes, ask Google, and others for help, and you learn from your mistakes.

A next step in learning data science might be learning to write Python or R. You might research data manipulation and data visualization tools, such as Pandas or ggplot. You are constantly making mistakes and learning from them, while also constantly refining your data science process. You might eventually turn to a new book, a MOOC (they still exist), or Udemy and Udacity. 

The process is not linear. The process is anything but linear.

In short, data science is a dynamic and evolving field - and your learning journey is likely to resemble that dynamism. To be successful in data science you need to always be learning, experimenting, and iterating - not just following a linear set of steps. So if you're on the data science learning journey, be sure to keep an open mind and embrace all that comes with it!

GitHub (Alone) Is Not A Portfolio

A second common misstep is: Using GitHub as it it were your portfolio. Don't hear what I'm not saying. You almost always will need a portfolio if you plan to transition into data science. If you don't actually need it - having one will help you. 

The creators of Git and then later GitHub did not design these version control systems as portfolio websites. Many use it as a portfolio website - but doing so is often a mistake. Instead the solution is to use GitHub as a back-end for a distributed portfolio. So you just need one. Please don't overlook the importance of this. 

The mistake is posting all of your work on GitHub and only on GitHub. Adding insult to injury it is also a mistake to just dump .ipynb files there. None of these quick hacks will serve you well. Instead it is important to build a distributed portfolio strategy. Here is what that means.

  • Focus on a few quality projects.
  • Make multiple portfolio entries based on that project (articles, blogs, videos, repos, etc.).
  • Multiple formats are key. Different formats will appeal differently to different folks. This is why multiple formats based on the same project are okay.
  • Distribute (keyword) those entries across multiple platforms.
  • Take your time. This doesn't happen overnight.

Here are four places (other than GitHub) you can put your distributed portfolio entries:

  1. On Medium (which would also be Towards Data Science, Towards Ai, or any of many other Medium - based publications).
  2. Just ordinary social media posts (Twitter, LinkedIn, Facebook, TicToc, etc.).
  3. YouTube... you don't have to have a zillion followers to use YouTube pragmatically. Have fun with it.
    Your own blog.
  4. Or, a LinkedIn Article (which is like a blog but on LinkedIn).

Reminder regarding GitHub. Let Github be a "backend" of sorts that the Medium articles, blogs, videos, and other entries link back to.

Pragmatic Points On Learning Data Science 

As promised there are a few more pragmatic points to review before concluding. It is also important to discuss A) setting realistic goals, B) staying organized, and C) asking for help when needed. 

A) Setting realistic goals. When you first learn data science, it's easy to set goals that are simply not realistic for where you currently stand in terms of experience or skill level. It is important to be honest with yourself about your current skills and limitations, and then set appropriate goals for the future. If you need help setting some goals - reach out to me and I'll help. I'm easy to find on social media.

B) Staying organized. Another common problem that data science learners face is a disorganized approach to their work. This can be due to several different factors, such as trying to do too much at once, or struggling with an over-complicated data analysis process. To stay organized when learning data science, it is important to keep track of what data you have collected, which tools you have used, and how your analysis is progressing. Another barrier to organization is that there are too many tools to choose from. One week you might work in Colab. Another week you might work locally with Anaconda. And the next week you might just work in another IDE. Balance the need to learn multiple tools with the need to stay organized in one simple tool.

C) Asking for help when needed. Finally, data science is a complex field, and it can sometimes be difficult to figure out how to get started with a particular data analysis project, or where you might be going wrong in your data science learning journey. To avoid getting stuck on these types of issues, always remember that you can reach out to other data scientists for advice, support and guidance. With the right mindset and approach, there's no limit to what you can achieve in data science! Here is an article on how to ask for help - and I also provided a webinar here on the same topic a while back.

Conclusion

Data science is a complex field, but with the right mindset and approach it can be incredibly rewarding. To help you on your data science learning journey this blog provided two major kinds of advice. The first kind was big picture. Specifically data science is a rapidly growing field that requires hands-on practice, mistakes, and iteration. Many data scientists make the mistake of trying to learn data science as if it were a linear process like they would have in school, which can lead to confusion and frustration. You have to embrace the mistakes.

I also wrote that "big picture speaking" you should also avoid using GitHub as their only portfolio platform and instead use it as a back-end for a distributed portfolio. You have to create multiple portfolio entries for each project and then place those entires across multiple platforms. I also reviewed a variety of platforms you should consider when looking to expand the range of platforms you use in your distributed portfolio entry.

The second major kind of advice I covered above was more pragmatic and tactical. To get the most out of data science learning journey, it's important to set realistic goals; stay organized by using one simple tool; and ask for help when needed. With these tips in mind, data scientists should feel empowered to explore this fascinating field without fear of making mistakes or getting stuck on challenging projects. Whether you're just starting your data science education or looking for ways to improve existing skills, remember that success lies in understanding how our brains process information - so don't forget about neuroscience sales tips! Keep up the hard work and enjoy all the successes data science has to offer!

If you have questions just reach out to me!

Now Offering Live Free Online Data Science Lessons.

 

Get You're Free Lesson Here