Chapter 11 How to Learn

In the last lesson we walked through a few interesting data science projects. Eventually, using the foundational skills learned, with practice on your own, and with other skills you pick up along the way, you’ll be completing your own, equally-awesome data science projects.

However, what many people don’t tell you early on is that that path will be paved with a lot of failure. This isn’t a bad thing! Data scientists fail all the time. They write code that produces an error they have to figure out. And they regularly have to abandon projects that aren’t going to work out. Failure is part of the process.

Failure

Even when a project is successful, know that there was failure on the way to success! The problem is that what you see in a final blog post or a product put out by data scientists at a company is the final product. This product may be something that is functional, really important, or even beautiful. What you don’t see is all the failure that happened on the way to getting the end product. Data science projects can be a lot like social media accounts. On social media, it’s easy to only show the good stuff about one’s life. For data science projects, the end product of a data science project may be awesome, so the user will only see the good stuff. But, there’s a lot of struggle and failure that went into creating the awesome end product!

success requires failure

In fact, that pathway to success in data science is always full of failure. And, often, failure followed by figuring out why you just failed is a great way to learn.

That doesn’t make failure easier. It will be frustrating from time to time, and figuring out why something isn’t working can be hard. That’s ok! Know that you’re not alone. Even experienced data scientists who have built really cool stuff experience lots of failure along the way.

process can be difficult

11.0.1 Learning How To Learn

In addition to learning the basics of data science in this course set, we also want you to learn how to learn.

First and foremost, the best way to learn data science is by doing it. Throughout these lessons, copy the code you see in the lessons and try it out on your own. If you get an error, that’s ok! Google that error and try to learn from this error! In fact, we’ve got a whole lesson in this course on how to Google and a lesson in a later chapter on how to get help for questions when you’re programming. But, there’s more to learning how to learn than getting good at searching on the Internet (although, that is important!)

11.0.1.1 The Mindset

Your mindset is very important to learning how to learn. Your goal should be to answer an interesting question. Your objective is not to memorize a bunch of functions. It’s to use those functions to do something interesting. The path to accomplishing that goal may be circuitous. You may take a few steps backward and experience a setback or two before moving forward. That’s ok!

mindset

11.0.1.2 The Path

When carrying out a data science projects, there is always more than one way to solve a problem. Your path may be different than someone else’s path.

In fact, while you may not know R code yet, the following four lines of code all produce the exact same output:

# Example 1
mtcars %>% tidyr::gather(key = variable, value = value)

# Example 2
tidyr::gather(mtcars, key = variable, value = value)

# Example 3
mtcars %>% tidyr::gather(key = variable)

# Example 4
mtcars_long <- tidyr::gather(mtcars, key = variable)

Any one of these would be a reasonable approach. We use this example to explain that there is more than one way to approach and to answer a question! Your path may be different than someone else’s. Your approaches may not be identical. And, that is more than ok!

path

11.0.1.3 Asking For Help

Data science is best done as a part of a community! Asking each other questions and brainstorming with others is ideally how data science projects are completed. In fact, this is how most companies work. You are part of a team of data scientists working and learning together.

Working together and asking questions:

Makes projects turn out better!
Helps everyone learn more!
Helps everyone have a better, less frustrating time!

No one knows everything and that is okay and expected.

We’ll point out where to find help when you’re stuck throughout this course set. However, it may not be obvious when to ask for help. While this is not a hard and fast rule, if you’ve been trying to find the answer to something you’re stuck on for half an hour and cannot figure it out, it may be time to post your question online for someone else to answer or to reach out directly to someone to get your question answered. During the half hour when you’re trying on your own, you should:

Google for the answer with various different wording attempts.
If it’s a coding question, you should try running code to test to see if the fixes from Google fix your problem.
If you’re getting error messages, paste those messages into Google and search.
If after trying all of these things you’re still stuck, then you should ask for help every time.

Rather than give up or get overly frustrated because you’re stuck, ask questions!

Ask Questions