You Ask, I Answer: Marketing Data Science Technical Skills?

In today's episode, Jessica asks, when it comes to marketing data science, I've got very good business knowledge, but lack of the technical side. Any advice? So there's a couple of questions to unpack here. Remember that in data science, data science is all about the intersection of four skills: business skills, technical skills, uh, mathematical skills, and scientific skills. And to be a data scientist, you have to have aptitudes in all of them.

You don't necessarily have to be expert, but you've got to be competent in all those areas so that you know what you're doing, why you're doing it, how you're gonna do it, uh, and what to do it with, what's what tech tools and techniques you'll need. The technical side is the part that people spend an awful lot of time on. Do I need to do, you know, be doing R or Python, uh, should I be using SQL servers? Should I be using uh you know NoSQL environments, what should the what are the tools? And for data science in particular, I would actually say that the mathematical and statistical side is more important to start with because the tools are how you do something.

The the math is what to do and and why you would choose a specific technique. Even something as really simple as the difference between average and median is important because they have different uses, use cases. You would use an average, for example, on a normal distribution, you would use a median on a non-normal distribution, like a power law curve. And so knowing the techniques will then guide you as to which technical skill you need. Uh what functions, you know, and which language will use.

If you go about it the wrong way, well, it's not wrong, but if you go about the technical first, mathematical second, you'll find yourself getting caught up in shiny object syndrome, and B choosing techniques that may not be appropriate for the problem you're trying to solve. So when it comes to Jessica's question in particular, the first question is: do you need the hands-on technical skills? You may not. Depending on your company, depending on the environment you're working in, if you have data scientists and such on staff already, you may be able to get help with the technical and mathematical, and you need to be versed in what's possible, what the right choices are, the uh the understanding the theoretical, the frameworks. You may not necessarily need to get your hands out and start writing code if you have people who can help do that, or you have an agency that will help you do that, but you need to know what to ask for.

It's like it's the difference between going to a restaurant and knowing what you want off the menu and in what order they go, versus going into the kitchen and cooking it yourself, right? You can have somebody else cook it for you, but you still need to tell them what you want and whether it's available and a good choice or not. Somebody's the waiter says, you know, what you want if you're appetizing you're like Boston cream pie, uh there'll be situations where that's not necessarily appropriate. Um the same is true in data science. So that's the first question.

Do you need the technical skills, or do you just need the know-how of what's possible so that you can orchestrate uh the project as more of a project manager? And then the second question really is that assessment of what skills do you have. Do you have the mathematical and statistical background? If you don't, again, I firmly believe that you'll be a better data scientist in the long run if you are versed in the statistical first and then in the technical second. Take a course.

There's a gajillion and a half courses out there in a you know in the are recording this in the middle of 2020. Um a whole bunch of them are free right now, and will probably be free for a little while longer. So go and take courses that will help you get the knowledge that you want, right? On the t on the statistical side. Then go take some technology courses.

Again, many, many, many for free out there. The big fork in the road that you're going to run into is going to be on the technical side. There's really two major languages, R and Python. Python, you see used a bit more on the machine learning side, R you see a bit more used on the statistical and data science side. But both are very capable.

Both are great at covering 96, 97% of use cases out there. And there are packages in each language that allow you to interoperate in other in the other language. Uh R has a package called Reticulate, which allows you to use Python code inside of R. Notebook environments like the Jupiter Notebook environment allow you to uh run multiple languages simultaneously as long as you know them and you can manipulate data in them. And so there's a lot that can you can do in those environments to interoperate.

So pick whichever one works better with your brain, because they are very different languages from a syntax perspective. Um start with that. I personally lean towards R. I'm a little older, you know, because there's all the gray hair hair here. Um and I grew up in languages like C and Java, which are more restrictive languages, so R feels more natural to me.

There are no shortage of people, including my kids, for whom Python is a lot more natural. It's it's easier for them. And I can hack around in Python, but it still is not as intuitive to me as R. Either way, the languages themselves are secondary to learn to how to think as a programmer. One of the resources that I think is actually a really powerful and useful resource is uh of a fun game language called Scratch by MIT.

If you go to Scratch.mit.edu, it lets you play around with little colored blocks that help you understand the concepts of programming with again without having to write code. And that kind of environment uh really gets you thinking about the what am I doing and why am I doing it, not necessarily the how of the implementation. So it's a really good intro to programming as a whole, and then you can use that knowledge and graduate to things like uh IBM Watson Studio, for example, which has the SPSS modeler inside, which is again those little colored blocks that you drag and drop in and connect them in sequence. If you if you're thinking about how to program and you learn in an environment like Scratch, you can graduate to a a professional environment and do the same things again without having to learn how to code. Um so when it comes to marketing data science, learn the statistical, then learn the technical, and on the technical side, choose a path, dabble around it first, see which one's more comfortable, take like an intro to R and an intro to Python, see which one feels better to you.

If neither feels better, you know that you're gonna be on the project management route because you're not you're not going to enjoy the technology. One of the things, and we'll talk about this in soft skills at some point, is that if you don't love the doing it part, don't make yourself do it. Yes, the salaries and data science are great and all that stuff, but if it doesn't make you happy, you're not going to do a good job, and you'll be miserable. Um, I don't particularly enjoy doing finance. I'm not good at it.

So it's it's uh you know you pick if you're going to to consider this as a career option, make sure you love doing it. Make sure that you want to do it. Uh you have follow-up questions, leave them in the comments box below. Subscribe to the YouTube channel on the newsletter. I'll talk to you soon.

Want help solving your company's data analytics and digital marketing problems? Visit TrustInsights.ai today and let us know how we can help you.