At Dataquest, our lessons are all built around answering real-world data science questions using real-world data. You’ll learn machine learning by predicting airline accidents, and you’ll learn statistics by finding the best players in the NBA.

Most courses also end with a guided project that asks you to synthesize the coding concepts you’ve just learned to solve data science problems just like a working data scientist would.

Working on these real-world projects is more interesting than answering practice problems or working with made-up data sets, and that helps keep students interested and motivated to learn. But there’s another advantage, too: building a portfolio of great data science projects is one of the best ways you can find a job and advance your data science career.

In fact, when we spoke to data science recruiters and hiring managers all over the world, we heard the same thing over and over again: data science portfolios and Github pages are among the first things they look at. Employers want to see if you can really do the job you’re being hired for, so having real-world projects to prove your skills you’re claiming on your resume is a must, particularly if you don’t have a fancy degree.

So while our projects are powerful learning tools that will keep you engaged and motivated as you work through our courses, they’re also a resource you can turn back to when it’s time to start the job hunt to showcase your skills to prospective employers.

One of the problems with online learning can be figuring out how to handle the sheer quantity of information that’s out there. Let’s say, for example, that you want to learn data science using Python. You should probably start with the basics of Python, and there are dozens if not hundreds of online courses out there that’ll teach you basic Python syntax.

But then what? Should you study some statistics next? Maybe try to learn `pandas` or `numpy`? Dive straight into machine learning? Focus on doing data viz? Whatever you choose, you may find that your next course requires some learning that wasn’t covered in your last course. Sifting through all of these choices and figuring out all the prerequisites is difficult and exhausting.

That’s why we created Paths: to help students go from zero to job-ready without having to waste time figuring out what they need to learn next.

Our data science courses are organized into four paths: Data Analyst in Python, Data Analyst in R, Data Scientist in Python, and Data Engineer. Each path is a sequence of courses, carefully arranged so that each course builds on the previous one. You won’t encounter prerequisites you don’t meet or have to search outside of our site for what to learn next. You can move smoothly from one course to the next and be confident that your learning is on the right track.

Of course, if you prefer a more a-la-carte learning style, that’s fine! The paths are totally optional and you can take all of our courses in any sequence you’d like. But many students find that the paths help them focus on actually learning and doing data science, since they don’t have to waste any time searching for courses, trying to figure out what skills they need, or Googling around to fill in prerequisites.

Video lectures are the foundation of a lot of online courses, but to be frank, we don’t think they work very well.

First, with videos, there’s too much of a time gap between learning and doing. We find that most students learn best when they’re presented with a concept and allowed to apply and experiment with it immediately. That’s why our platform features a split browser window, with information presented on the left side and a live code environment on the right/ That way, there’s no gap at all between learning how to do something and actually doing it.

Second, videos waste too much of your time. When you’re asked to watch a 30-minute lecture before you put your hands on the keyboard and start coding, it’s inevitable that you’re going to miss or forget things. That means that when you start coding and hit a roadblock, you have to waste a lot of time scrubbing back and forth across the video or scanning through the transcript, trying to find the part that’s relevant. We think searching through text is faster, and with our platform’s search feature, it shouldn’t take more than a second or two to find exactly what you’re looking for.

Third, doing data science in the real world requires English-language reading skills. When you’re working as a data scientist, the documentation for that API you need to use or that package you’d like to learn is going to be available in text format, not as video lectures. So while you’re reading through our text-based lessons, you’re also getting practice reading code documentation and getting exposed to the terms and phrases you’re going to see in the documentation as a working professional.

Finally, we’ve found that videos can trick you into feeling like you learned something without actually teaching you. We’ve all had the experience of watching somebody else write and explain their code and thinking “That looks easy enough, I get it,” and then sitting down to reproduce it later and realizing it didn’t sink in as well as we thought. That’s the reason our courses are broken down into smaller lessons and then even smaller screens: we’ve found that going hands-on and applying each new concept as you learn it helps you retain the information better.

And of course, there’s plenty of science to suggest that the lecture format in general doesn’t work particularly well, whether it’s in video format or lectures in real life.