3 Pitfalls to Avoid When Learning Data Science Online
Online learning has been around since the late 1990s. MIT OpenCourseWare (CW) led the way by publishing educational materials for most of the undergraduate and graduate classes at MIT. This includes videos of recorded lectures, the slides the professor used in class, and PDF's of the problem and solution sets.While the first phase of online learning was about increasing access to content through the web, the second phase of learning aims to take advantage of what's unique about the medium of the web. Unlike previous media platforms like books, radio, and television, the web has a few key properties:
- The web is a responsive medium (people can have different experiences from each other)
- The web facilitates real-time communication with others
- Computers can simulate all older media (text, audio, video, etc.)
Good online learning experiences "take advantage of what the web has to offer." With this guiding principle in mind, we'll discuss the following pitfalls to avoid when learning online.
- Passive over Active Learning
- Certificates Over a Portfolio of Projects
- Learning Without A Community
Passive Over Active Learning
Passive learning usually involves consuming static content like audio, video, and text. While consuming static content can explain a concept well and bring you into the right context, it's a poor way to build deeper understanding. Active learning, on the other hand, encourages learners to:
- Apply the concept they've just learned to a different scenario
- Reconcile their pre-existing knowledge with what the new concept has taught them
A key component to making active learning work well is feedback. Feedback is critical to understanding if you're making progress or not towards your learning or career goals. When choosing where to learn, look for the following feedback opportunities:
- Automatic feedback on highly structured exercises (multiple choice, code answer checking, etc.)
- Subjective feedback on projects from peers or mentors
- High level feedback on your progression in learning (ideally from a mentor)
Certificates Over a Portfolio of Projects
Certificates are very popular in online learning, but are often implemented very poorly. A certificate is any way that showcases your expertise and signals to others (usually employers) that you're a capable person. A good certificate is one shows mastery and is difficult to achieve. If you can easily receive a certificate without much effort, than 'so can everyone else'. A degree from a top university is still a good certificate, because of the:
- Difficulty of getting accepted (Stanford's acceptance rate in 2016 was 4.8%)
- Difficulty of finishing the coursework
- Scarcity of that degree (Stanford graduates about 3,000 to 4,000 students every year)
- Brand of that top university (built up over centuries)
Certificates do have a place in online learning, but look for learning programs that:
- Focus on helping you learn depth and mastery (achieving a shallow understanding doesn't make you compelling in the marketplace)
- Help you build a portfolio of projects that showcase that mastery you've achieved
- Help you meet your learning and career goals (whether that's learning for fun or switching careers)
A good portfolio of projects that augments the story you want to tell can help you a lot more than a generic credential. In the following example, a hypothetical student was able to build a portfolio project in her online learning program that aligned with her previous experience in healthcare:
Learning Without A Community
Learning online can be a lonely process and it's helpful to be a part of a community of other learners. Having a learning community you can participate in can help you:
- clear up any misunderstandings you may have about a concept
- solidify concepts you've learned by helping others clear up their misunderstandings/li>
- find other learners to collaborate on projects
When evaluating online learning providers, pay close attention to the type of learning community they foster. Learning communities often exist at a few different levels of depth:
- A forum or Q&A thread attached to a specific lesson or course; this is useful for very specific questions, at the time of learning
- A semi-structured live chat community, hosted by a tool like Slack; this is useful for longer, more free-form discussions to build better understanding of a concept, share career stories, or find other members to collaborate with.
How We Avoid These Pitfalls
Dataquest is an online platform for learning data science and data engineering. I lead the content team and we think a lot about how to improve the learning experience online. Here's how we avoid these pitfalls:
Passive over active learning
We use text, custom diagrams, and code examples to explain and introduce concepts. We then ask you to immediately complete a code exercise that uses that concept. You receive feedback immediately every time you submit your code. At the end of each course, we've created "Guided Projects", which let you practice synthesizing concepts learned across multiple lessons in new ways.
Certificates over a portfolio of projects
While we do award a certificate when you complete a course or track, we focus a lot more on building a portfolio of interesting projects. We even have a dedicated plan where we provide feedback on these projects.
Learning without a community
We're proud to have a very active Slack community where learners help each other get unstuck and collaborate on projects. The learning community is mentioned in many of the reviews on SwitchUp.
Other Inspiring Examples
Here are two other online learning providers that think about these pitfalls as well:
Khan Academy - While Khan Academy does use a lot of videos, they provide ample opportunities for practicing in most of their courses using quizzes and tests.
Codeacademy - One of the original inspirations for Dataquest and the most popular learn-to-code platform out there. Like Dataquest, Codecademy focuses heavily on learning by completing coding exercises and building projects.
What has your experience been with online learning? We'd love to learn the hurdles you've run into while learning online @dataquestio.