Image by story set on Freepik

Let me start by acknowledging that this topic has been written and discussed many times on many different platforms. So what can I suggest that hasn’t already been said?

Well, as it turns out, I’m a lot older than the average average author. Although some people politely call me “mature” or “experienced”, the truth is that my perspective is shaped by the fact that I’ve been around long enough to remember when neon colors and gradients on bar charts were cool . So what I can offer you is advice that has stood the test of time.

As an analyst, you will probably spend at least 10 times more time analyzing the data than you will have to present the information to your audience. Therefore, graphs and other visual representations are important because they create a better understanding even for those who are not trained to analyze data sets. By building visualizations, you can help your company’s decision makers understand complex ideas at a glance.

So how do you improve this skill? Well, there are many learning opportunities in the form of online courses and many tools specialized in visualization. However, I hesitate to recommend a specific technology or course because the world is changing so quickly and I want to offer advice that stands the test of time.

So make it a habit to research other people’s work. Create a bookmarks folder called “Inspiration” and fill it with blog links. Dedicate 15 minutes a week to browsing blogs to fill your brain with possibilities.

I also recommend buying a few books. Edward Tuft is the grandfather of data visualization, and I’m also a fan of Nathan Yau at, where you can find books, blogs, courses, and tutorials. I keep both of his books, Data points and visualize this, on my inspiration shelf.

The mechanics of how you actually create the visualization will come with practice. Some of you will fall in love with R or python and others with Excel or Tableau. The key is to keep your brain full of possibilities so you can imagine what you want before you build it.

Data Scientists are famous for saying that “80% of building a machine learning model is preparing and cleaning the data”. However, this also applies to data analysts. Whatever the actual percentage, the truth is that much of the time spent with data, in general, is spent cleaning it.

Data cleaning is important because raw data can create misleading patterns and lead to wrong conclusions. Early in my career, I was tasked with delivering the percentage of support calls that were resolved on the first try. As I briefed my management on the results, one of them insisted that the numbers “didn’t make sense.” When I researched more, I found that call data has a column called “status” that is sometimes populated with an “X”. Obviously, the system will record an “X” for a test record that should be ignored.

Data cleansing skills grow with hands-on practice and business experience. This makes acceleration difficult. Generally sites like or Zindi are not the best places to practice data cleaning because they are focused on data science, and the datasets are usually pretty clean already. On the other hand, government websites such as or are a great place to find datasets that are messed up. You can also follow Tidy Tuesday project, even if you’re not an R user, to find interesting datasets and learn about the kinds of cleanup steps that occur in nature.

As I mentioned, I’m hesitant to recommend a particular tool or technology because of how quickly the landscape is changing. It’s probably safe to say that Python and R are here to stay, but SQL is on another level. SQL is the language of databases; therefore, learning SQL will always be the most direct way you can manipulate and study datasets for your analysis.

If you work for a company that allows you to pull statements into Excel spreadsheets – ask if it’s possible to access the database directly with SQL. Once you become familiar with database structures and writing SQL to get the data presented the way you want, your efficiency will increase along with the quality of your work.

There are many resources available to help you up your SQL game, CodeAcademy, Udemyand Udacity are great free resources for finding hands-on courses. SQL generator and SQL beautifier are great links to keep handy and will help you learn. Stack overflow has a great community to answer technical SQL questions if you get stuck.

Here at Rasgo, we’re heavily invested in giving back to the data community. Our most recent project was the launch of our free SQL generator which generates the SQL syntax required for specific data transformations. We found that people were searching Google and Stack Overflow for the required SQL syntax – wasting a lot of time that could be used for data analysis. SQL Generator is an SQL query template that lets you customize the column names and table structure, choose the operation you want to perform, and then constructs the syntax for you in many different “flavors” of SQL. Never stress over the subtle differences between DATEDIFF() and DATE_DIFF() again! Read it this post from earlier this month, for some other useful online tools.

Critical thinking is one of the hardest things to learn because there aren’t many courses or a one-size-fits-all approach to mastering this skill. Critical thinking is a conscious effort to challenge the automatic mental processes that rule over us.

Critical thinking is not something you naturally possess or something that once acquired is always “on”. Instead, critical thinking is something you consciously activate when you have an important decision or analysis. It is a purposeful effort to challenge, question, and confirm hypotheses.

While there are many different resources available to help you improve your critical thinking skills, the one I want to focus on is asking questions. If you get into the habit of writing down a lot of questions, you will begin to think critically.

Let’s pretend you’ve been asked to look at customer churn and whether it’s been improving lately. Stop asking yourself questions like:

  • How is the outflow calculated? Why so? Are there other ways?
  • What is churn from the customer’s perspective
  • Is it always the customer’s choice?
  • What factors might matter? Can I measure them?
  • What is the purpose of this task? Prove, disprove, demonstrate, support?
  • What could I have overlooked? Are there known assumptions that I can’t confirm/prove?

Notice how this line of questioning forms layers upon layers. Answering a question can lead to more questions. That’s perfectly fine. That’s the point!

For example, when I was an analyst at a telecommunications company, I was asked to look at potential reasons why churn had suddenly increased in the previous month. (it was April at the moment). I used critical thinking by writing a list of questions similar to the list above.

When I got to the question of how ebb is calculated, I challenged myself to think if there were other ways to measure ebb. In short, I found that:

  • The company uses month-end numbers to calculate churn
  • The month of March was actually a “fiscal” month that covered February
  • February was the shortest month of the year
  • Using an alternative ebb formula showed that the percentage did no increase in March

Basically, the low tide formula was such that it would “jump” in the shortest month of the year, which was called March, simply because of how the math works.

The moral of the story is that critical thinking sometimes leads you to conclusions you wouldn’t expect. This is also something you can decide to improve today; it’s not some mythical soft-skill you’re born with!

Communication on the other hand, is soft skill. You may be the most talented and insightful data analyst the world has ever seen, but it won’t help if you can’t communicate with others. And by others, unfortunately, I mean non-technical people.

As an analyst, you are in two different worlds. In one world, you need to address technical questions with your peers and other data experts. In the other world, you are a translator for business-oriented decision makers. You need to provide clear, high-level explanations in a supportive, not confusing, way.

One way to start practicing this soft skill is to get involved in a community. There are many online communities, websites and forums for data analysts to join. I’m a big fan of slime and discord channels as a way to interact with others, but they’ve been known to die out quickly. i found Locally optimistic and DataTalks.Club be a good consistent place for analysts to exit. You should also practice your communication skills by starting a blog. Average is a great place to start where you can exercise your creativity and practice, practice, practice.

It’s still important to keep up with the latest trends and acquire skills in things that make you marketable. For example, 20 years ago VBA, macros, DAX and ASP were great skills to give you an edge in promotion. These days, those skills are less likely to matter. Hopefully I have given you some useful advice on the skills that they are not changed over the past 20 years, so you can avoid getting lost in the weeds. If you ever want to contact me and talk about the old days, you can find me hanging out Locally optimistic and DataTalks.Cluband you can always email me directly at Rasgo.

Josh Berry (@Twitter) leads Customer Facing Data Science at Rasgo and has been in the data and analytics profession since 2008. Josh spent 10 years at Comcast, where he built the data science team and was a key owner of Comcast’s internally developed feature store – one of the first stores with features to come to market. After Comcast, Josh was a critical leader in building Customer Facing Data Science at DataRobot. In his spare time, Josh performs sophisticated analysis on interesting topics such as baseball, Formula 1 racing, housing market forecasts, and more.

Original. Republished with permission.

Previous articleI Won’t Sell My Soul to Modern Medicine: Treating a Physician’s Moral Injury
Next articleApple TV Plus Baseball: How to watch stream Red Sox vs. Yankees, Cardinals vs. Dodgers for free on Friday