Can I Learn Data Science On My Own?

You can teach yourself to do just about anything, with the possible exception of making tamales, for which you do require another on-site human being to demonstrate how to shape, fill, and wrap the little bundles of cornmeal masa. According to that theory, you can certainly teach yourself data science if you have the patience, determination, and interest to see it through. There are on-demand classes, free tutorials on YouTube, and, yes, even books on the subject, from which you can learn a great deal. Whether it’s the most advisable way to learn the subject if your goal is to use it professionally is, however, another matter. But it certainly can be done.

Advantages to Self-Teaching Data Science

Teaching yourself data science gives you complete control over the learning process: you learn what you want, when you want, at the pace you want, and to the endpoint of your choosing. No other teaching modality can offer you that. If you are an autodidact by temperament, there may well be no more enjoyable way to learn the highways and byways of data science. 

If you’re the head of a small business and want to harness the capabilities of data to aid your organization, self-teaching may be a good approach to the subject. You won’t have to show up for a class at any set time, and you’ll be able to study when your duties don’t take you elsewhere. You needn’t become an all-around expert, as no one’s going to test you on your data science knowledge, and you can tailor your curriculum to the bits that will be of the most use to you. And, that way, you won’t have to hire an expensive data scientist or analyst, at least not until your data work has paid off, yours is no longer a very small company, and you can afford a data science department of your very own. Bonus: you’ll be able to understand the work your data scientists are doing, and not depend entirely on the visualizations they create for the benefit of laypeople. Certainly, brushing up on your linear algebra while running a business is going to be a challenge, but it’s possible.

The other considerable advantage of self-teaching is that the cost is far more reasonable than that of a live class with a human instructor. Although you can’t quite do it for free, and the most respected series of books on the subject will cost you approximately $45.00 each on Amazon, you’re still looking at hundreds of dollars rather than thousands of them. If you’re not sure you want to make the sizable investment a bootcamp involves, trying the field on for size by teaching yourself can be an excellent way of scoping out the terrain and making sure your brain is compatible with data science. Your outlay needn’t be that large, and it may save you a small fortune in tuition for a class you won’t enjoy.

Data Science Self-Teaching Tools

You’ll have a lot of options from which to choose if you decide to try to teach yourself data science. You’ll probably begin by consulting the platform to which people turn to learn how to reset a circuit breaker or to get their dogs to stop barking: YouTube. There’s quite a selection of data science tutorials there. One lasts five minutes, another lasts ten hours, and one goes on for nearly 20 hours—and you have your pick of all of them. Do note that the ten-hour video is designed to get you to purchase an advanced course; the 20-hour one is entirely free.

A second option is an on-demand course. In these classes, you’ll be watching video tutorials, rather like the ones on YouTube, although you have the assurance that they’ve been curated by a school and are very likely more reliable than whatever you manage to find for free. You can watch these whenever you like, at whatever pace you like, although there is often a time limit of something like six months on how long you have to complete the course. These classes aren’t quite self-teaching in the strictest sense of the term, but they may well appeal to all but the most rigorous and misanthropic of autodidacts.

Perhaps this last group would fare best by buying the book, or, rather, several books, as data science is a complex field that has to be approached from a variety of angles. You need to learn math, you need to learn Python, you need to learn database technology (SQL and NoSQL), you need to learn machine learning, and you need to know how to create data visualizations to communicate your conclusions to the people in charge. You can start out with Data Science for Dummies (Amazon offers the tome for $26.00), and follow that with the O’Reilly series of books on the subject, which makes an excellent (if rather steeply priced) resource.

Drawbacks to Learning Data Science on Your Own

All that said, learning data science on your own isn’t free of drawbacks. The biggest of these is inherent in teaching yourself anything: the absence of someone to whom you can ask a question when you fail to understand something. A living, breathing, and otherwise sentient resource like that is invaluable. There’s no sense in pretending that data science isn’t a complex subject to learn: a lot of complicated concepts are involved, and you need to master everything from basic programming to machine learning, and from statistics to database querying languages and technologies. You’re going to run into a roadblock somewhere in there (it’s just in the nature of the beast.) Asking for help is probably anathema to you if you’re one of those counterdependent people who like doing things on their own, but sometimes you have no choice. You can only go in circles in the market looking for Marshmallow Fluff so many times before it gets silly, and you finally break down and ask one of the nice people who work there whether the Fluff is hidden on the ice cream toppings endcap or in the baking aisle.

A further shortcoming of teaching yourself a workplace skill such as data science is that when you’re finished, and no matter how thorough a job you may have done in teaching yourself, you’ll have nothing in the way of a certificate or a diploma to attest to your newfound abilities. Sure, the hiring process for any data science job is going to include a technical interview, but a number of obstacles stands between you and the chance to demonstrate your skills, some of which are going to be insuperable for a resume that lacks some sort of proof that you know what you say you know. Without that, the odds of your making it past an automated application tracking system are slim indeed.

Alternatives to Learning Data Science on Your Own

If the above seems to be pointing you in the direction that, no matter how badly you want to teach yourself anything, you’re better off (at least in the twin cases of data science and tamales) with a live instructor, and, as the live instructor’s natural habitat is a live class, in some kind of school setup that puts students on one side of the room facing an instructor armed with a dry-erase marker. That’s the learning paradigm you started to experience in kindergarten; frustration with it may well be what turned you into an autodidact in the first place.

Classes with live teachers come in two varieties. The first is the live in-person class that puts you, the teacher, and your fellow students in the same room in a brick-and-mortar school building. The alternative is the live online class that uses the magic of the internet to put you in one location in space, your instructor in another, and your fellow students in their own, while you all work together at the same instant in time.

The former is, of course, the kindergarten model described above, and requires commuting to class, sitting in a room with people who chatter, text under the table, wear too much cologne, and ask annoying questions about what the teacher just explained while you’re trying to understand what the teacher is explaining now. The other problem you’ll encounter with live in-person data science classes is that they’re about as rare as astatine in all but the largest markets, which means New York, Los Angeles, and Chicago

Of course, you can always register for a degree program at a local college, university, or community college, but that’s going to be a lengthy undertaking, and perhaps not your cup of tea if you initially wanted to learn data science on your own. You may not have the time, the funds, or the desire to sit in class for four years to learn how to forecast how much tamale masa a market needs to prepare for each of the days leading up to Christmas.

Accelerated, career-oriented classes these days tend to be broadcast using the internet to reach participants anywhere on the globe. This type of class has a number of advantages over the brick-and-mortar-school ones, not the least of which is that you don’t have to spend time, money, gasoline, and patience to get to a class, most of which require traveling during rush hour. You also get to study from your own space, in a seat that is most certainly going to be more comfortable than a standard-issue classroom chair with one leg ever so slightly shorter than the other three

You’re also not cut off from the instructor, even if you are a face in a box on a screen. There are multiple ways of signaling for the teacher’s attention (including just raising your hand), and the teacher will answer any questions you may have right there on the spot. If you get your fingers all tangled up in a practical exercise, you can also invite the teacher to take a gander at your screen and help untangle you as necessary. A live instructor will also keep track of your progress, and, very importantly, provide a rooting section to get you across the course’s finish line.

Examples of the online, all-in-one, complete certificate programs available from Noble Desktop include its Data Science Certificate program, which will provide you with a solid grounding in most aspects of data science (going light on the linear algebra.) You can complete the course in not much longer than a month. If you wish to be more thorough, you can consider the Data Analytics Certificate program: it includes everything you get in the data science bootcamp, plus modules on using Excel for data analysis and Tableau to create comely visualizations and dashboards to allow your stakeholders to grasp what your analyses are all about. Noble Desktop offers briefer courses as well, which might pair well with some on-your-own study.

Noble Desktop is one of several schools that include with their certificate programs a further human resource in the form of mentoring sessions. These sessions, which put you into 1-to-1 live contact with an experienced data science professional, give you a chance to ask any further questions you might have about material covered in class or about data science in general. Your mentor can also assist you with preparing your resume, cover letters, and portfolio for your assault on the job market at the end of your course.

All these benefits of a live class simply aren’t available to you if you elect to study data science on your own. If you’re just curious about the subject, you can probably manage well enough by buying the book(s), but, if you have serious professional aspirations, you’re going to be infinitely better off in a class with a human teacher. And you’ll emerge from your program with a certificate of completion that will vouch for you and your newly-acquired abilities. If you want to be taken seriously by HR officers (and that’s the goal of most people who take classes to better their professional lot), that certificate is going to be worth quite a bit more than its digital weight in Bitcoin.