This is FiveMinuteFriday, Data Analyst Versus Data Scientist.
Welcome back to the SuperDataScience podcast everybody, super excited to have you back here on the show. And as you may have heard from the podcast lately, we’ve been hiring, we’ve been growing our team, and one of the positions that I am super excited about that I actually just published the job description for, or the position opening for is called a data analyst. It took us some time, but finally we’re hiring a data analyst.
And in this podcast, in today’s episode, I wanted to talk about the difference between data analyst and data scientist, because I really had to sit down and think, “Who do we need on the team right now for this specific role? Do we need a data analyst or data scientist?” And that allowed me to refine better in my mind the difference between the two and come to a concrete answer. And the reason why we’re hiring a data analyst is because of this difference.
So let’s talk about this for a bit. Now you may have heard me talk about the analytic value escalator by Gartner. I also referenced it in my book, Confident Data Skills, because I believe it’s a great tool for understanding the different processes in analytics and different level of maturity of analytics. So the analytic value escalator by Gartner says that there are four main stages or four main types of analytics, and each one gets more and more complex as you progress. So there’s descriptive analytics, there’s diagnostic analytics, there’s predictive analytics, and there’s prescriptive analytics.
Descriptive analytics is in the past. This is describing what happened. You’re just visualizing the data, trying to decide what happened. Then there’s diagnostic analytics, and that is answering your question, why did it happen? So it’s still in the past, it’s answering a question about something that happened, but just a bit differently. It’s taking it the step forward. Not just what happened, here we go, let’s have a look, but it’s also like, let’s think about it. Why did it happen? What were the reasons for it? What are the triggers? Then the next stage is predictive analytics. That’s in the future. It answers the question, what will happen? And that’s where you can apply a model that your trained on past data to make a forecast in the future, or to… Just basically you train a model and then you run it in production in the future to get results all the time. So that’s what will happen. That’s predictive analytics. For instance, who will buy a product versus who will not buy a product. Who will click on the link, who will not click on the link. Which vehicles likely fail and need maintenance and which vehicles will likely not fail and so on.
And finally prescriptive analytics is, what action do you prescribe? According to Gartner it answers the question, how can we make it happen? I would like to rephrase that. I would say, what can we do from that? What are our actions? What should we do based on… Maybe we don’t want it to happen. Maybe we want the opposite to happen. Maybe we want to prevent it from happening, like vehicle failures and things like that. So there we go. So we go from the past with descriptive and diagnostic analytics, to predictive analytics and prescriptive analytics in the future. And this is a great starting point for what is the difference between a data analyst and data scientist. In my mind, my opinion on it, you will find lots of different definitions and comparisons. But in my mind, it’s very simple. A data analyst does descriptive and to some extent, diagnostic analytics.
So the main duty, the main responsibility of a data analyst is to take the data from past, as I mentioned on the podcast probably before. Data is the past. It’s the role of a data analyst is to take data, which is describing the past, and describe it and to put into visualizations, put it into… Maybe apply some basic machine learning like clustering techniques or classification techniques, or a linear regression, but not for the purpose of predictive analysis, but for the purpose of descriptive analytics. To describe what happened. So there’s a lot of visualization, a lot of Tableau, Power BI and stuff like that. And some basic machine learning models but for the purposes of descriptive analytics. So that’s the data analyst. And also answering questions, why did it happen. Helping us with those questions, working with stakeholders, I think there’s probably a difference between data analyst and a senior data analyst. A senior data analyst would be more involved in this diagnostic analytics, trying to understand why did it happen, So that’s the data analyst or senior data analyst.
And then when you move on to data scientists, data scientists should be able to do all those same things, but also go venture into the world of predictive analytics and predict when it’ll happen. That’s where all the hardcore machine learning comes in. Artificial intelligence, deep learning, all these approaches and techniques, methodologies, tools. That’s where the tools really started growing. That’s where the main core difference between data scientist and data analysts lies. Data scientists doesn’t just describe what happened, describe the past try and answer the question, why did it happen? But also answer question when it will happen in the future or when it will not happen in the future? How can we build a model around this? How can we predict the future? How can we do a forecast and things like that. So they look into the future more.
And also then prescriptive analytics. So that’s another role of data scientists. Not only, “Hey, when will it happen?” That’s a very junior data scientist, so what will happen in the future, but a data scientist should also be able to answer the question, how can we influence this? What can we do? What are the actions? There’s little point in insights from data scientist if you can’t put actions towards them. For business, a business needs to be able to take actions, whether you’re describing the past or the future, there’s no point in those insights if you’re not going to take any actions. You’re not writing a history book. You you’re actually running a business. A business is running, something needs to be happening.
So if you’re a data analyst, you describe what happened, and then the executives and managers make their own decisions what to do. But as a data scientist, you create the models and then you prescribe, what should we be doing and why in order to make this happen or make something else happen, and what actions should we take? So that’s a data scientist. And having that my mind is quite easy to make the decision right now, in this team, or in this specific business unit, we need a data analyst because we’re going to be running a lot of experiments. We’re going to be running experiments with a SuperDataScience membership product and improving it, changing it, adding different kinds of features and things like that, and we need somebody who’s able to analyze, “Okay, how did our users respond to that? And why, perhaps, did they respond to that?”
And of course it should be beyond the simple AB test. There’ll be new products. The person needs to be able to understand how to collect data, how to set up data points, how to analyze it and all these steps is the data science process. So it’s not a simple AB test, like why did that… Which button do they prefer? It’s much more involved than that. But at the same time, we don’t need somebody to make predictions or how will they behave? How will users behave in the future? What features should be better or worse? We’re hiring another person for that, we’re hiring a product person who’s going to approach from a creative side of things, a slightly different approach. So we don’t need a data scientist to make predictions about future, but we do need a data analyst to analyze the data and describe what it means and help us understand why or help the product manager understand why did that happen.
So there you go, that’s the main difference between a data scientist and data analyst. To recap, a data analyst looks into the past and describes what happened there, perhaps answers the question why it happened. Data scientists does all the same things, but also looks into the future and predicts what will happen and what actions we need to take to make more of it happen or less of it happen? That’s the distinction, again, according to my opinion, between the two. And if you’re interested in a role in a data analyst and got a few more sentences to say on that, the careers that are at www.superdatascience.com/careers. I know we have a lot of people who are experienced with Tableau who are listening to this. You don’t have to be just experienced. Specifically in the job description I said you don’t necessarily need experience, but you need either experience or a demonstrated portfolio of things that you’ve done and really cool things that you’ve performed.
So I just want to say a heads up. I know there’s probably going to be a lot of people applying for this position. We’re looking for somebody who can give us the wow effect. If you’ve got some really cool data sets you’ve worked with, really interesting, you can show that you’re actually passionate about visualizing stuff, about getting insights from data, about extracting information about what happened in a very efficient way, communicated really well. That’s who we’re looking for. So make sure that your portfolio has outstanding visualizations. Something that we can go and browse through and be like, “Wow, we really got to talk to this person. You’re definitely on top of the game.” Of course, Tableau certifications will be a plus, a huge plus, if you have your Tableau certification. Once again, positions up at www.superdatascience.com/careers. Have a look if you’re interested. And on that note, I look forward to seeing you back here next time. Until then, happy analyzing.