3 Steps to learn Data Analysis for non-analysts

Anastasia
5 min readOct 20, 2020
Photo by Scott Graham on Unsplash

Imagine you have to put up that cool presentation for the customer to sell your product and you need to get in there some fancy numbers. The problem is your Analytics team is 30% on vacation, 30% in some conference and the rest has a backlog as deep and dark as the Mariana Trench. All the unlucky requests which are wandering around at the bottom of it look like those crazy fishes which never see the light of day — all those fancy DS projects with will probably never be done.

So, you’re in a little bit of a pickle now. You have to magically make use of the data you know is available somewhere or escalate the request to the DS team manager and they might prioritize it.

Good luck!

Data analysis is going to be in every part of the organization. Most of the tech companies are democratizing data now and the rest will follow.

They want all their employees to be comfortable with extracting, analyzing, visualizing, and making decisions based on data.

Everyone will have to work with data one way or another. Take the first step now to be ahead of that trend.

Step 1: Build a relationship with your analytics team.

This is the easiest, most rewarding, and straightforward way to get to know the data and the data people.

If you have a data/product analytics team in your company — talk to them! They’re going to appreciate it.

Find out where do they get the data from, how is it stored, what are the ways to get access, and how strict is it. What are their favorite tools and visualization libraries? What are the dashboards they can’t start the day without? What is their favorite way to brew coffee? Yes, all that matters for networking.

Essentially, your main goal is to find out those three things:

  1. What kind of data is available, where and how it is stored, and how can you get access to it if possible and need be?
  2. What are their tools of choice and how can you start using them even a little bit?
  3. How do they build their backlog and prioritize their projects and tasks?

Step 2: Embed that analytical thinking in your life.

Data is all around us. Every interaction with a device, app, the product is tracked, stored, and made sense of and it’s time to understand how it can be used for our advantage.

Analytical and data-centered thinking is the key to success in any career.

Ask questions, read those data collection/protection legal documents when you sign up for services, download your data from Facebook, Twitter, Tinder, etc. Know what kind of data could be collected and try to figure out uses for it.

The more you learn, the easier it is to get into the analytical mindset.

A fair warning, if you are just starting out it can be overwhelming and quite paranoia-inducing to find out how much data is stored and recorded about you. Don’t fall victim to your own curiosity. Be grounded and maybe do remove some of the privacy settings you use for some of the apps. 🤞

When you think about how data can be useful for you to improve the product you are working on or your personal life, ask yourself those 3 questions:

  1. Why? What is the goal to improve? There should be a reason for a question, even though sometimes they can be just for fun. The best use of analytics is to answer the questions you care about.
  2. What? What data could you find available? Whether it’s doing research on the datasets existing in your company or requesting your own data from Facebook or Google, knowing what kind of data and in what granularity you can have to work with is key.
  3. How? How can you use it? Is it sufficient? Is it correct? Is it enough? Can you rely on the results?

I also recommend taking courses in analytical and critical thinking. They help to make sense of data and to not fall victim to biases and false assumptions. Being able to correctly and objectively interpret data is crucial. Here’s one from Coursera I liked: link.

Step 3. Learn the tools.

Ask for the training, learn on your own, and bonus point for those who code — learn some visualization libraries and/or build your own tools.

1. Spreadsheets

Data Scientists tend to be a bit snobbish about spreadsheets, but they keep immense power and can do so much more. Learn the fundamentals of MS Excel (here) or Google Sheets. If you have never worked with them, start with your own example, use your own data. For example, I have a spreadsheet where I track all my travels and spending on them in different categories, and I can see the value for myself in it, so I am invested in exploring it and making it better.

2. Visualizations

Data is often used to support a statement. A story. Words seem to hold more value when there is a good looking convincing chart to back them up. (Unless it’s Softbank’s presentations, check out here and here. It’s hilarious.)

Misleading charts are used to dramatize the effects seen in data and lead the reader to believe their side of the story.
Here, instead of showing that the sales for 3 products are somewhat similar, the chart on the left dramatizes the reader’s perception and makes it look like Product 2 is selling much better. Although, that’s not the reality.

A few very useful resources I found on data visualization are Google data visualization guidelines here and here, and a couple of blogs on the rules of visualization: here and here.

3. Coding

Even though Python or R is not the holy grail of data science, some coding skills definitely give you an advantage in how fast and flexible you are in manipulating data.

The most valuable coding language for data manipulation is SQL.

You’d be surprised how flexible is it in providing access to data and being able to do fairly complex analytical operations. As a bonus point, Analysts at your work would appreciate you knowing how to do proper joins, and Here’s a nice course on SQL for data science here.

And when it comes to choosing between Python or R, if you are not planning to work as a Data Scientist, I recommend going with whatever feels best for you. You could use React, JavaScript, or even C++ to parse and visualize data. Where there is a will there’s a way.

Check out my video, where I discuss this topic in more detail.

--

--

Anastasia

Data Scientist who believes everyone should know how to use and analyse data. Working in fintech, living in Stockholm.