A guide to learning data analysis skills
This is the introductory post for an in-depth series on what data analysis skills if you want to start, specialize or transition into data analysis.
- Don’t focus solely on technical skills, but on translating data into business value.
- Learn tool-agnostic skills.
- Have a baseline proficiency in spreadsheets, basic stats, and basic data viz. Also, learn how to connect data analysis with business decisions.
- If you want to further go down the data analysis route add SQL, dashboarding, and data storytelling to this.
- Want to become an expert data analyst? Learn Python and ETL tools.
Not sure which data analysis skills to learn? You are not alone.
A question I frequently see and get is what skills you should learn when you want to get started with data analysis. Especially when you don’t have a technical background in math, statistics, or computer science and want to transition into data analysis from another role it’s hard to understand what skills you should learn.
Usually, when this question gets asked in a forum the answers mention seemingly random skills without any clear guidance on what is important and in which order you should learn the skills.
Similar to many other in-demand professional fields data analysis is a field that’s been ravaged by spammy information and bad ebooks, and it’s difficult to know where to start and what’s worth learning.
So, wouldn’t it be great if there would be a clear guide or framework showing the exact learning path you should follow when it comes to data analysis? Detailing what you should know when it comes to the different skills? Specifically for a marketing context?
A couple of years ago I was in a similar position when I wanted to transition from being a marketing consultant into data analysis. I was overwhelmed by all the skills, courses, and paths I could learn and take.
Based on that experience and being in the field now for a couple of years I developed a skill development framework that gives you clear guidance on what skills to learn, how in-depth, and in which order when you want to get started with data analysis.
Three misconceptions when it comes to data analysis skills
Before we dive into the framework and talk about what data analysis skills you should learn exactly there are usually three misconceptions that come up among the answers when somebody asks what data analysis skills they should learn:
First, the recommendations focus too much on technical skills
E.g. “learn PowerBI, SQL, R, etc.”
The best technical skills don’t matter if you can’t translate your data analysis into business value, i.e. business decisions. You did your analysis in Google Sheets but saved the company hundreds of thousands in costs? Well, congrats! Nobody (and I mean nobody) in upper management will ask what cool technologies and coding libraries you used.
And this is also your biggest strength as somebody with a marketing/business background: You have strong business acumen. You know what matters for the business and what provides value. Play that strength by positioning yourself as somebody in between the data and the business world. That’s how you can compete easily against those with a purely technical background, by providing business value (that’s what I did when I transitioned). And that’s also why I focus so much on teaching a data analysis framework that gives you a structured framework for translating data into business outcomes.
Second, the recommended skills focus on specific tools
E.g. “learn Tableau, Google Analytics, etc.”
Yes, those are probably some of the most used tools out there. However, chances are that at every job you will have at least some tools for website analytics, dashboards, or whatever that are not mainstream.
So rather than learning the ins-and-outs of a specific tool (also on the job, you can still always Google any challenge you encounter…) learn agnostic skills which you can transfer to different tools.
For example, instead of knowing every detailed feature in Tableau, rather learn what makes a good dashboard, how it should be structured, and how it provides value to the audience (relating directly back to what I wrote above). Rather treat Tableau as a vehicle to learn (and show) those things than learning it for its own sake.
This makes you a lot more flexible when it comes to looking for new job opportunities and you don’t risk getting stuck when a new market leader among the tools emerges.
Third, too advanced skills are recommended for beginners
E.g. “learn Python, R, etc.”
I bet a dollar to a doughnut that every time somebody asks in a community “I want to transition into data analysis. What should I learn?” there will be at least one answer saying “Python”. Often followed by “If you want to be taken seriously in data analysis spreadsheets and SQL won’t be enough”.
Do you want to hear an industry secret? It’s not true. Even seasoned analysts often won’t use Python that much. Rather SQL and spreadsheets are often the bread and butter of many analysts and data analytics organizations. Of course, Python has many unique advantages and a rightful place in any technology professional’s toolkit for certain specialized advanced use cases. But more often than not it’s used by individuals and organizations to appear “advanced” and more “glamorous”.
This ties in directly with what I said above about the first misconception: If your analysis provides business value, nobody cares which technologies you used. Don’t start at the top and begin learning Python in a course just for the sake of it. Instead of wasting time learning theoretical concepts, think about how you can apply your analytical skills the fastest. Start with the fundamental skills and use them already to drive actual value for the business (and as such to build an analytical portfolio for yourself).
Now, let’s get finally started with the learning framework!
The Data Analysis Skill Pyramid
The Data Analysis Skill Pyramid is a framework to give you a clear learning path on what to learn when it comes to data analysis – without wasting time on nonessential things. It will help you to decide which data analysis skills to prioritize depending on your goals: No matter if you are a marketer who wants a future-proof skill to work more data-driven and to stand out among peers or if you want to transition into data analysis as a career path.
It’s a system that you can use to teach yourself data analysis in only a few months by focusing on the right things.
How the framework is structured
All the skills in the pyramid are ordered along two axes:
The horizontal axis shows you if a skill is rather a soft skill or a technical skill. For some skills, there won’t be a clear distinction between soft and technical as they provide communication tools (e.g. data visualization). However, in general, the more to the right as skill is placed the more it is a soft skill, and the more to the left it is a technical skill.
The vertical axis gives some guidance on how much scale and quality you will be able to provide with the respective data analysis skills. Scale means in this context automating things, reaching larger audiences, and being able to work with larger data sets. Depth means that you can access more data and do more sophisticated analysis which you couldn’t with lower-level skills. More on that down below.
The pyramid is divided into three stages representing beginner, intermediate, and advanced data analysis skills respectively. Below gives you an overview of what each stage stands for.
Step 1: Beginner skills (One-off analysis)
This stage is all about doing simpler to intermediate one-off analysis without a lot of automation. At this stage, you should focus on learning how to translate business challenges into data analysis projects, i.e. how to turn data into actionable insights into business decisions.
Learn a data analysis framework and process to connect your data analysis with business value and to focus on the right data and which data analysis methods to use. This teaches you what to do and you can start learning the how.
As such become a spreadsheets wizard and know the ins and outs of Excel or Google Sheets (doesn’t matter which one). Have some understanding of applied stats and modeling. (You don’t need a Ph.D. here. Again, applicability > in-depth theoretical knowledge.) And know some basic data visualization to communicate your analysis findings (e.g. in spreadsheets or on slides).
Including some general business acumen, those five skills are it to get a foot in the door and to have a basic set of data analysis skills.
Don’t get me wrong. You can already have a tremendous impact by knowing them. Again, in the end, it doesn’t matter what technologies or techniques you use, the only thing that matters is the value your analysis creates for the business. You saved your business $100k by doing some marketing channel optimization analysis in a spreadsheet? Great! Nobody will care if you used Google Sheets for it…
Step 2: Intermediate skills (Improving impact)
This stage is about working more independently, autonomously, and having more impact.
You basically have three options to improve your impact with your data analysis:
- You can use more data that you didn’t have access to before (and/or do more advanced analysis with it).
- You can reach a larger audience with your analysis.
- You convince other stakeholders better with your data analysis.
The skills in this stage will help you with all three things.
SQL will help you to get more data autonomously (and the developers will thank you for not taking up their time anymore). As such it will help you to be more flexible with the kind of data sources you can use and how you can manipulate the data you pull. Also, you can start automating first things.
Dashboards are the natural successor to the basic data visualization of stage one. They allow you to reach large audiences at scale vs. the one-off charts you might create with slides or sheets.
Data Storytelling is an important one in any business or marketing context. If you want to have an impact you have to convince other people. Nothing better for this than an awesome presentation or pitch. This is a catalysator for the insights you generate with more technical skills. Again, you can have a tremendous impact no matter the technology you use if your outcome is directly linked to business objectives. Data well presented in a well-structured and action-oriented slide deck provides you with the leverage for this.
Step 3: Advanced skills (Automation)
I wasn’t completely honest with you, there is actually a fourth option to having more impact with data analysis: Doing more of it.
And the best way of doing this by automating things. ETL tools refer to tools that extract, transform, load data from one or more data sources into an output that represents the data differently from the source or in a different context than the source. ETL tools let you build data pipelines that allow ongoing data analysis. They’ll automatically pull data from data sources, format it, store it somewhere, and give you output such as dashboards.
With coding languages such as Python or R, you can automate things even further or do more complex analyses such as with machine learning.
So, should you learn all the skills in the framework?
It really depends on your goals. If you want to become an expert data analyst then probably yes. If you are a marketer who wants to stand out among peers, then you probably won’t need all the skills from the top of the pyramid. Either way, no matter the goal: Start at the bottom and work towards the top.
To sum this up, let’s work through an example on how a typical learning path could look like:
Let’s say you are a marketer and want to learn some data analysis skills. What should you learn?
If you are completely new to data analysis start at the bottom. Master stage one to be able to do a one-off analysis. If you want to become a better data (analysis) driven marketer you should make sure that you can check off all the skills listed in that stage.
As such as a marketer you should already have some decent business acumen. Continue with learning a data analysis framework to have a bullet-proof process to make your analysis valuable and to translate your analysis into business decisions. Afterward continue with basic data viz, applied stats, and spreadsheets. Ideally, you learn those at the same time and in a marketing context to make sure you don’t only learn theoretical concepts but know how to apply them in practical use cases right away. Again you learn everything as a means to an end to provide business value rather than for its own sake.
Only after you have mastered stage one continue on your learning path towards stage two. As I believe data storytelling is so important as it is a catalysator and leverage for everything else you do when it comes to the more technical aspects of data analysis I recommend starting with that one before continuing with SQL.
Are the skills of stage three useful for a marketer? Yes, of course. However, are they necessary? Probably not…
The skills from stage three have some diminishing returns when considering investing time there. Consider those skills either when you want to become a data analyst or later on in your learning path if you want to specialize or dive deeper into data analysis. As you can see it really depends on your goals here. Either way, if you are just starting don’t make learning Python your priority for now.
The next parts of this series go more in-depth with the individual skills in each stage. As such we will start with stage one in the next part and I’ll detail what exactly you should know when it comes to beginner data analysis skills to get the fundamentals in that stage covered.
I’ll send out the next parts of the series via email first. Sign-up below to not miss out!