SDS 178: Visualization in Data Science

SDS 178: Visualization in Data Science

Welcome to the FiveMinuteFriday Episode of the Super Data Science Podcast!

Today, we’ll be talking about the importance of data visualization and Tableau, one of the best tools out there. This episode will be a sneak peek also to two courses that we’re about to relaunch soon so make sure to tune in!

If you’ve been following Super Data Science and my courses online for quite some time now, chances are you’ve seen the emails floating around regarding the revamp of my two courses. I’m talking about the Tableau 2018 Basic and Tableau 2018 Advanced Courses.

Today’s episode is inspired by the two courses. We will be talking about data visualization. Yes, this is very related to the previous FMF episode about storytelling. If you’ve not listened to it yet, you can check it out here. We’ve talked about how big the effect of storytelling is when we want our audience to easily grasp any kind of information. And to make storytelling more effective, we have to properly utilize the reach of visualization.

Visualization has two purposes: for data mining and for presentation.

Visualization can be used for data mining. Most of time, data scientists handle huge amount of data making it very difficult to go through them. It’s hard to explore and search for clues. Where do you start building your models? You want to understand first what’s important and what’s not important. When we have visuals, it becomes clearer and easier. You see everything. You get the feel of your data. We are able to transform it to something we can understand.

Visualization can be used for presentation. It is at the very end of the data cycle. Again, people retain information much better if we use storytelling rather than throw hard facts at them. Many parts of the brain get easily triggered if we tell them through visual and compelling narrative. We should seek visualization that not just convey insights but engages the audience, make them feel, make them fascinated, make them say ‘wow’, etc.

The demand is high for data scientists and the supply is high. But what’s the missing link? Data scientists should be able to convey insights and make it engaging for the audience. Bridge the gap.

If this episode got you curious and interested about data visualization, make sure to check out the new courses about Tableau. I’m quite excited about this huge step so I hope you guys are as well!

Items mentioned in the podcast:



Full Podcast Transcript

Expand to view full transcript

Hey, guys. Welcome to the Super Data Science Podcast. This is FiveMinuteFriday, and today, we're talking about the importance of visualization in data science.

Today, I'm at the Queensland University of Technology Library. Queensland University of Technology or QUT for short is a wonderful university in the city of Brisbane in the state of Queensland in Australia. Today, we're talking about visualization, so you probably have already noticed a couple of emails floating around and a few invitations for you to come and join us in the revamped Tableau 2018 course.

In fact, we're launching two courses at the same time this week. We're launching the Tableau Basic 2018 Course and the Tableau 2018 Advanced Course at the same time, so we'd love to see you in there, and today's Five-Minute Friday is inspired by those courses to talk about the topic of visualization.

Last week, during the Five-Minute Friday, we spoke about the importance of storytelling in data science. Well, visualization is something that can help with storytelling. In fact, visualization has two components. This is my personal view. Visualization in data science can be used for two distinct purposes, and we're going to actually start off a different purpose, not the one that's related to storytelling, but like ... It is in a way, but it isn't.
The first purpose of visualization that I define for myself is visualization for data mining. When you have a lot of data, when you have tons, and tons of rows and columns, and data points in your dataset, it can be very difficult through them, very difficult to understand what's going on and make sense of it even for yourself, even to understand where to start building your model to explore the data to perform that crucial step of data mining when you're looking for clues, when you're looking for answer.

We're not even ready yet to build a model because you want to understand, what are the important features? What are the not important features? What anomalies exist in your dataset? What does your dataset look like, and what does it feel like? If we were machines, we would just like look through all that data, all those tables and spreadsheets, whatever other data we have, and we would understand right away what's going on. We'd be able to, from there, start our analytics, but we are not.

For us as humans, it's really hard to look at a spreadsheet of a billion roles and try to get something from it. What we can do on the other hand is we can look at visuals, and that's where visualization comes in in this first purpose where it allows you and helps you to see your data, to get a feel for your data, to transform your data from just numbers into something that can actually ... that you can interact with, that you can understand as the explorer, as the investigator.

That's why I really like visualization a lot. At the very start of my projects, before I even create any models, even before ... like even sometimes before I do the proper data progression when I can just like get the data, throw it into the visualization tool, solve this like a drag-and-drop fast tool. I already might be able to get some insight from it. It might even help me during the data progression phase to understand, "Okay. What are the anomalies? All right. What do the distributions look like and other things like that?" That's data visualization for data mining.

On the other hand, we have visualization for presentation, and that ties in very well with what we talked about last time on the previous week's Five-Minute Friday where we're talking about the importance of storytelling. As you remember, when we just throw hard facts at people, only two areas of the brain are actually active. When we tell a story, we can engage way more areas of the brain, and that will help with compression of it and retention of information.

That's why it's so important to be able to come up with not just the visualization that conveys the insights, but a visualization that actually tells a story, that engages your audience, that gets them intrigued, that gets them fascinated, that makes them say, "Wow," that makes them ask more questions and get the answers from that same visualization or the next visualization in your slide deck in your presentation.

That is, my friends, the importance and value of visualization. In fact, we currently are in a state of data science where there's lots of companies demanding data scientists and there's lots of data scientists that are ready to provide these services. This field has been around for a couple of years now, and it's slowly starting to get to the phase where the demand is high and the supply is high, but what is actually the missing link are those data scientists that can not only get the insights, but actually convey the insights to the audience in a way that's going to be engaging, that's going to be fun, that's going to help them understand.

Those are the top, top data scientists, and one of the most important skills that they have is visualizing the data or the insights that they have and presenting them to the audience, so that is the importance of visualization. It can help you in two phases. It can help you at the very start of your data science project, and it can help you at the very end of your data science project in order to make you one of the most successful data scientists.

If you're excited about that and you've seen some of our Tableau invitations for Tableau 2018 courses that we're releasing this week, then jump on board. Tableau is an amazing tool, one of my favorite. In fact, my favorite tool for visualizing data. It's very drag-and-drop tool. Very easy to use, and we'd love to help you master that tool and become a master visualizer. On that note, thank you for being here today, and I can't wait to see you back here next time. Until then, happy analyzing.

Kirill Eremenko
Kirill Eremenko

I’m a Data Scientist and Entrepreneur. I also teach Data Science Online and host the SDS podcast where I interview some of the most inspiring Data Scientists from all around the world. I am passionate about bringing Data Science and Analytics to the world!

What are you waiting for?


as seen on: