This is a guest post written by Dataquest, an interactive online data science learning platform.
There’s no shortage of discussion in the data science community about where to best learn data science. However, there’s precious little discussion of a topic that’s arguably more important: how to learn data science.
Taking the right approach to learning matters. For all the differences between programs and platforms, the reality is that you get out of any educational experience what you put into it. Approaching your data science learning journey with the right mindset and the right game plan will help you get the most out of your studies, and make you a more effective data scientist in the long run.
So how should you learn data science? Here are three crucial things you need to do no matter what platform, bootcamp, university, or educational program you’re learning from.
3 Keys for Learning Data Science
Work on Personal Projects
Learning all of the technical skills associated with data science is a lengthy process, and it’s one that you’re unlikely to reach the end of unless you find an effective way to keep yourself motivated and integrate your interests into your learning.
The easiest way to do this is to find some free time for personal data science projects interspersed throughout your studies. If you’re passionate about climate change, find a unique angle for analysis and dive into some climate data in your free time. If you love soccer, find a stats site, learn to scrape it, and start working on an analysis of your favorite players.
The key is to pick topics for these projects that genuinely interest you. Find a question related to your interest that you’re curious about, and then set out to use data to answer it. That way, you’ll feel motivated to keep working even when the actual work you’re doing at the moment isn’t particularly thrilling.
Why is this important? Even if you love working with data, there are aspects of learning data science that are likely to frustrate or bore you. For example, if you dislike data cleaning — a critical but not particularly fun data science skill — it can be challenging to motivate yourself to get much practice with it. If you’re working on a personal project you care about, trying to answer a question that really interests you, it’s easier to motivate yourself on those days when you don’t feel like practicing just for the sake of practice.
Working on personal projects in your free time also has the pleasant side effect of getting you ready for the process of job applications. If you don’t have any work experience, your applications for entry-level data jobs are going to ride mostly on the strength of the projects you’ve done. If you’ve been working on personal projects throughout your studies, you should reach the beginning of the job application process with a portfolio of projects that are thoughtful and unique. This may save you some time having to prepare new projects, and it will also prevent you from applying to jobs with the same five “homework” projects everyone else in your class has on their GitHub, too.
Apply What You’ve Learned Frequently
Study after study has shown that students who apply what they’re learning fail at significantly lower rates than students who do not. It is critically important, then, that wherever you’re learning data science, you’re also taking the time to apply it as you learn.
This can be a tipping point for some data science students, particularly if your course of study is primarily lecture-based. It’s easy to watch a video lecture and feel like you’ve understood the material, especially if the presenter is a good teacher. But understanding something on an intellectual level, and being able to apply it in the real world, is not the same thing. Data scientists need to be able to do both.
Working on personal projects will certainly help you apply what you’ve learned, but if your learning platform doesn’t integrate more frequent, shorter hands-on sessions, then you’ll want to make sure you’re getting that critical practice yourself. If you don’t practice applying concepts quickly after learning them, you may find that by the time you get to the relevant section of your personal project, you’ve already forgotten what you learned.
For example, if you’ve just watched a video lecture on For Loops in Python, you should follow that by opening up a Jupyter Notebook of your own, importing some data, and writing some For Loops. Ideally, you should practice applying a concept directly after learning it, and then several more times throughout the week to ensure that you’ve cemented how to apply it into your long-term memory.
Stay Engaged with Peers and Mentors
It’s important to make interaction and communication a part of your data science study. It’s easy to get technical tunnel vision and focus on tweaking your algorithms until they’re as accurate as possible, but in real-world data science work, building a great model is only half the battle. Your highly-accurate model will only be useful if you’re a skilled communicator who can explain what it means to others, and convince the higher-ups at your company to act on your results. Working with peers and mentors as you study data science will help you learn how to talk about these topics effectively and convincingly.
Finding a mentor has other benefits, of course — a good mentor will help keep you on the right track, and point out areas for improvement you might not be able to see on your own. They also can often help you make important connections and assist you in your job search, once you reach that stage of your studies.
Working with peers is important too, though. Teaching a concept to a peer is one of the most effective ways to test whether you truly understand something, and working together with other students on data science projects will give you experience working as part of a data science team, and help you practice workflow-related data science skills like using Git and GitHub effectively for collaboration.
How you engage with peers and mentors will probably depend, quite a bit, on how you’re studying. If you’re enrolled in a bootcamp or a university program, this kind of interaction has probably been arranged for you, but if you’re working on an online platform or doing self-study, you may have to be more proactive in seeking it out. Luckily, there are many online data science communities, and you should be able to find data science and/or programming meetups in most cities (if not, you can start one yourself!).
Don’t forget about social media, either — there are cool data science groups and communities on most major social media platforms, and if you get involved, you’re likely to make some useful connections as you’re interacting with and learning from the other people on the platform.
How to Study More Effectively
While those big-picture keys will help you be successful in your data science studies, there are also some smaller-scale things you can do to help ensure you’re learning at peak efficiency.
Make Clear, Explicit Plans (With Contingency Plans)
Studies like this one have shown that people are more likely to follow through on their plans when those plans are clear and specific. “I’m going to learn data science” is a pretty vague plan. “I’m going to study data science for five hours each week” is a little better. “I’m going to study data science at my desk from 8 pm to 11 pm every Tuesday and Thursday each week, and make up any session I have to miss on Saturday morning from 8 to 11 am” is better still.
Having a contingency plan as a back-up is particularly important because, in the long term, you will miss study sessions from time to time as things come up in your everyday life. If you don’t have a back-up plan, you’re less likely to actually make the work up.
Regardless of how you’re learning, note-taking is a worthwhile endeavor that will help you retain what you’ve learned. There is some evidence that writing out your notes longhand is better than typing them, but you’ll benefit from note-taking even on a computer so long as you:
- Don’t transcribe verbatim or copy-paste things. A big part of what makes note-taking effective is that you’re writing out what you’ve learned in your own words. If you copy-paste, you lose this cognitive benefit.
- Review your notes after taking them, and again at regular periods over time to keep them fresh in your mind.
- Test yourself against them. For example, cover up the “For Loops” section of your notes and see if you can remember the syntax, then check your notes to be sure you remembered correctly.
Leave Your Phone Somewhere Else
It doesn’t matter how disciplined you are. Studies like this one have demonstrated a “phone proximity effect” — your phone can impact your cognitive performance when it’s nearby, even when it’s out of sight and switched off! Even if you think you’re not being affected, you probably are — most respondents in the linked study said their phone’s proximity didn’t impact them, but their scores proved otherwise.
The lesson here? When you’re going into a study session, leave your phone somewhere far away. This may not always be practical, but when you do have the opportunity, it’s best to leave it switched off and in a different room, behind a closed door. Students who left their phones in a different room scored better on memory capacity and fluid intelligence tests than students who left their phones on their desks or in their pockets or handbags as they worked.
Studies suggest that there are biological reasons some people procrastinate more than others, but even serial procrastinators can improve their study habits by making a few key tweaks like:
- - Breaking down big tasks into smaller ones
- - Giving yourself rewards for completing tasks (and varying what those rewards are)
- - Setting clear deadlines (or having them set for you)
- - Do your best to keep things light and fun — procrastinators tend to procrastinate more when they perceive a task to be very important. In the context of learning data science, this is another reason why it’s important to work on projects that you’re genuinely interested in, as these tend to feel more fun.
- - Forgive yourself for mistakes. Forgiving yourself for past procrastination may make you less likely to do it again.
Get more science-based study tips from Dataquest here.