How to Avoid These 4 Time-Consuming Mistakes While Teaching Yourself Data Analysis
If you've ever attempted to self-teach yourself something, you know that it often takes you an immeasurably long amount of time. This rings especially true when it comes to teaching yourself Data Analysis.
From coding to mathematics to data visualization, data analysis encapsulates three very different skills that all require various methods of learning. Not only that, but it can be easy to get sucked down the rabbit hole of Youtube tutorials and data analysis articles which can lead you to even spend more time learning than you had originally anticipated.
I've made my fair share of mistakes while learning data analysis, the top of which I've listed below, along with some practical ways to avoid them that you can begin implementing immediately. Learning data analysis doesn't have to be a long process as long as you know when to put pen to paper (so to speak), when to use the tools you already know instead of learning new ones, when to focus on learning only the skills you need, and when to put your foot down and finish your goal of learning data analysis.
Add to an existing project every time you learn something new
The mistake: Sitting in coding tutorial-purgatory thinking that you know how to code without actually knowing how to code.
The fix: The simplest way to avoid wasting time in coding tutorial purgatory is to add to an existing project every time you learn something new.
Before studying software development at university, I attempted to teach myself to code many times by following online tutorials from Codecademy and Youtube. Unfortunately, I just spent all of my time doing the tutorials and never actually putting pen to paper (as it were) and creating something using what I learned. I wasted a lot of time not applying what I learned theoretically which resulted in a lack of tangible results.
Learning the coding aspect of data analysis is probably the most hands-on portion that requires immediate implementation of what you've practiced via a tutorial. For example, after writing out your whole analysis program, you could watch a tutorial on how to group your code into functions. Then, you should go into your project code and create functions that organize and simplify your code.
Not only does this practice allow you to learn more quickly, but it also allows you to complete projects more quickly. For example, you may only be learning data analysis to complete a single project. Instead of waiting to work on your project until you've learned everything there is about coding in Python, you could begin chipping away at your project along the way as you learn different skills or techniques.
For example, I need to learn certain aspects of data analysis to complete some undergraduate research that I'm undertaking. However, because there's a deadline for me to complete the research stint and because I need to graduate eventually, I need to work quickly to get a grasp of what I need to add to my code and then implement it immediately.
Remember: you don't need much depth of knowledge to complete a data analysis, you just need to have the breadth of knowledge to understand when you need to do something and when you shouldn't.
Use Excel for data cleaning (and anything else you can think of)
The mistake: Thinking I have to use a special tool to complete my data cleaning which means I spend less time cleaning and more time learning.
The fix: Use Excel as much as you can to complete any and all data analysis tasks.
Social media makes it appear that you need to be a wizard with a multitude of coding languages, tools, and platforms to become a data analyst. In fact, many data analysts do their jobs just fine by only using Excel. I'm not sure when Excel became uncool, but I'm here to say that you should use Excel for all of your data cleaning, and basically anything else you can think of.
Excel is still an incredibly powerful tool for working with data and it's one of the easiest to learn thanks to a multitude of resources online.
When I was first learning data analysis, I wasted so much time learning all these new tools, coding tricks, and database hacks to work with my data. Instead, I could have immediately begun cranking out clean data and even entire analyses if I had just started using Excel from the beginning.
The philosophy I want to impart to you is that, if you know a way of doing something and it works, keep using it. Yes, there may come a point in your career where you need to expand past using Excel, but for right now, don't waste time learning how to conduct data analyses – instead, do more data analysis with the tools you already know!
Focus on linear algebra, probability, and statistics
The mistake: Believing I needed to know multivariable calculus and discrete mathematics to conduct data analyses.
The fix: Only learn the math that you need to complete an analysis – chances are you already know everything that you need.
One of the biggest misconceptions about self-teaching data analysis is that you need to know advanced forms of mathematics. This is where terminology is important. Some people will either confuse or purposefully interchange the terms "Data Science" with "data analysis" and then base their description of the discipline on data science. In fact, the two are completely different. If you want to do data science, then yes, you will need to know advanced forms of mathematics. However, if you want to do data analysis, you can get away with much simpler forms of mathematics.
You can accomplish most data analysis tasks using linear algebra, probability, and statistics (with the potential for some single-variable calculus thrown in there).
By focusing on the math that you actually need to complete your analysis, you'll save time that would otherwise be wasted on self-studying tricky mathematical concepts that you may never use. The best way to do this is to look at all of the mathematical requirements for a project and only learn those that you're missing. This will shorten your learning time from potentially several months down to only a couple of weeks. Furthermore, this allows you to focus on your breadth of knowledge as opposed to your depth of knowledge. Remember, breadth of knowledge can be picked up quickly and implemented immediately. Depth of knowledge can take months or years to develop.
Make a learning plan with a deadline that focuses on the right goal
The mistake: Treating the learning part like an endless timeline instead of setting hard deadlines that produce concrete results.
The fix: Create a learning plan with a hard deadline that focuses on achieving a specific goal or a short list of objectives.
Remember when you were in school and the teacher would give you a syllabus at the beginning of the semester containing the learning objectives of the course and everything you would need to learn to achieve those objectives? Those were invaluable documents that should be the inspiration to help us create our data analysis learning plans.
You need to create your own syllabus that has a hard deadline and a clear objective. You don't want to get carried away with your learning only to realize three years later that you're not quite sure what you've just learned and why it may or may not be relevant.
For example, I'll be working on some undergraduate research that will require me to conduct a data analysis. The data analysis is only a small part of the project, which means that I need to create a learning schedule for the aspects that I need to self-teach that also takes into account all of the other parts of the project that I need to complete. In other words, I can't spend the whole time teaching myself data analysis concepts because there's a whole list of other things that need to get done. Furthermore, the learning plan needs to have a few clear objectives that, once achieved, are left behind so I can begin implementing the skills into my project. As I said previously, I don't want to go down a rabbit hole of learning only to realize that what I just learned wasn't necessarily a good use of my time.
The trick is to create a learning plan with a hard deadline and three clear objectives. These objectives should either refer to practical skills, add up to the completion of a project, or to a job in data analysis. The objectives must be this way so that once you've completed some learning, you can assess whether you've achieved the objective or whether you need to do some further learning. The goal is to give a definitive "yes" when asked if you could complete any of the objectives in a test-like scenario.
Not only does this method keep you on track with your learning but it also saves you from getting distracted, falling off course, studying the wrong things, or taking years to finish what could be completed in just a few weeks.
Subscribe to get my stories sent directly to your inbox: Story Subscription
Please become a member to get unlimited access to Medium using my referral link (I will receive a small commission at no extra cost to you): Medium Membership
Support my writing by donating to fund the creation of more stories like this one: Donate
