SDS 184: Why You Need Domain Knowledge in Data Science

SDS 184: Why You Need Domain Knowledge in Data Science

Domain Knowledge in Data Science
Welcome to another FiveMinuteFriday Episode of the SuperDataScience Podcast!

Today, I’ll be sharing to you a simple hack perfect for people who want to finally start, make the shift, or just simply want to be more valuable in an industry. So, start listening to learn the concepts of the FIFO and LIFO to easily acquire any domain knowledge you want!

Just like any file or folder you’re seeing on your desktop every time you open your computer, our skills and specialties as data scientists are transferable from one space to another – from one industry to another.

When needed, this skillset you’re offering as a data scientist should be flexible to wherever you are placed – may it be in healthcare, banking, real estate, transportation, security, etc. The number of places where data scientists could impact and excel is indeed beyond finite.

So, it’s already established that you’ve got the knowledge, skills, and specialties just like the others. But every industry will look for the same skillset. What will separate you from others?

It’s domain knowledge…

Certain industries require a different understanding of their data, models, and outputs. It’s essential that we gain the needed knowledge in the industry we’re going to so we make an impactful development. We don’t have to get a degree to learn; what we do is get the existing resources and learn from it. We ask ourselves: if they want this kind of outcome, what type of data you need? Or, if they used this kind of data, what models should you arrive to? And all that stuff.

We just have to find an efficient and effective system for us to acquire basic familiarity on the industry where we’re going. This is where the concepts of FIFO and LIFO comes in.

FIFO (First In, First Out). The first element that comes in is the first element that comes out. Examples: queue to your favorite sundae, warehousing of perishable items, customer service

LIFO (Last In, First Out). The last element that comes in is the first element that comes out. Examples: stacking books, washing dishes, unpacking your weekend bag

Getting the ‘feel’ of using FIFO and LIFO in your day-to-day or your career could go a long way. As data scientists, there are times that we don’t notice small details and intricacies to what is presented to us. We don’t see the implications it will have on the problem. There may be consequences that could greatly affect the business. And, there are opportunities and benefits we’re missing if we know nothing about how important FIFO and LIFO are in understanding every industry’s data lifecycle.

DID YOU ENJOY THE PODCAST?

  • Have you noticed anything around you where you can apply the concepts of FIFO or LIFO?
  • Download The Transcript
  • Music Credit: Lightning by Distrion & Alex Skrindo [NCS Release]

0

Full Podcast Transcript

Expand to view full transcript

This is Five Minute Friday, episode number 184, FIFO versus LIFO.

Hello and welcome back to the Super Data Science Podcast, ladies and gentlemen. Today, we've got a bit of a nerdy episode. I'll admit this one's quite nerdy. What is it about?

Sometimes we talk about the importance of domain knowledge in data science. That is true. Domain knowledge is extremely important. The technical aspects of our roles as data scientists and the skills are extremely transferrable. The algorithms that you are applying one day at a pharmaceutical company, you might be applying at a railway operations company and then, another day, you might be applying it at a space agency, and another day you might applying at a bank.

Those components of our roles are transferrable. We are lucky that way, that we can very quickly adapt to the environment. But that adaptation actually requires some effort, and that effort is to acquire the domain knowledge. And domain knowledge is what are the specifics of the pharmaceutical industry? What are the specifics of a railway company or the railway industry? What are the specifics of the space industry? And what are the specifics of the bank that you're working for now?

And so in order to acquire this domain knowledge, I've developed some tips and tricks and hacks over the years, and today I want to share one of them with you, and this one is called the FIFO versus LIFO. So what does it stand for? Well, if you are a programmer, like a hard core programmer, and you've worked with Assembler before, then you will know what I'm talking about. FIFO stands for First In, First Out. LIFO stands for Last In, First Out. And also, you will probably know what I'm talking about if you are an accountant, because this is important in accounting, as well.

And so what are these two concepts and why are they important? Well, these concepts are related to how things or items or assets or data is processed, or how it is dealt with in a process. And the names speak for themselves. First In, First Out means that if you have an element, whether it's some data or it's a physical asset. It might be a burger. It might be a car. It might be a person standing in queue. Basically, FIFO means an element that came first or entered a system first will be the first one to leave it. They leave in that order. So if you have elements A, B, C, and they entered in that order, A, B, C, then they will leave in that order, A, B, C. And we'll have a couple examples just now.

But before that, LIFO stands for the opposite, Last In, First Out. So if elements entered in the order A, B, C, they will leave in the order C, B, A. So let's have a look at a couple examples to make this a bit clearer, and then we'll understand how this is actually valuable in understanding domain knowledge and data science.

So examples of FIFO. Another word for FIFO is queue. And this is why. So for instance, people queuing up to buy movie tickets, that is an example of FIFO in action, because the first person in the queue will be the first person to get the ticket and will be the first person out the queue. The last person in the queue will be the last person to get their ticket and will be the last person out the queue. So hence, first in, first out.

Another example is warehousing of perishable items. So if, for instance, you have a grocery store that is warehousing bread, the truck comes and loads the bread. The bread that comes in first, or the batch that comes in first ... Let's say that there's a batch that came in first, Batch A, then two hours later or a day later, Batch B came in, and then Batch C came in another day later, well, Batch A is gonna be the first one out. Batch B is the second one out. Batch C is the third one out. Same thing goes for medicine. If medicine has been warehoused, they use a FIFO system to get it out of the warehouse, because it's perishable.

If they're using the opposite, a LIFO system, so Last In, First Out, then they've got a problem, because the medicine that came in first is gonna stay around for a very long time, and it's gonna expire. So that's not necessarily gonna work, in that case.

So you can see the domain knowledge coming in. So if you're in a grocery store or you're in a movie business or you're working as a data scientist for a medical company, a pharmaceutical company, you already can see the value of this.

All right, what's another example? Another example is a phone answering queue. If you're on the phone to a bank and they've got support representatives answering the phone, then again, it's first in, first out, because if it was the other way, then the person who called first would be on the phone forever. Another example is cars being loaded onto a two-sided ferry, or cars using a two-sided ferry. So a two-sided ferry is one of those that you have in Europe, that are crossing these rivers all the time. I'm sure there is in other parts of the world, as well. I was just recently on one myself. You enter on one side. Then your car goes to the end. Then it crosses the river and then the opposite side opens and then you drive out. So in those cases, usually the cars that come in first are let out first.

Whereas, if it's a ferry with one side, like a long distance ferry that goes on for 10 hours, usually they only have one entrance. So the cars get in. They get packed, and so basically, the cars that come in last are blocking the cars that came in first, so therefore, the cars that came in last will go out first. So that's a LIFO system. So you can see, even in ferries, there can be two types of approaches, FIFO versus LIFO.

So that brings us to LIFO. LIFO is Last In, First Out. We already had that example with the ferries. So let's have a look at a couple more. And by the way, another way of saying LIFO is STACK, because when you stack things up, like you stack up books onto a table, the first book that you stacked will be at the bottom. You cannot pick it out first, because the whole thing will tumble. So you have to pick up the top book first, so the book that you put in last. So a stack of books will be a LIFO example.

Another example of LIFO is cement bags. So when you're warehousing the cement bags, you're gonna be stacking them on top, and because it's not a perishable item, it's okay to keep the bottom bags stacked for ages, for years, perhaps. I'm not an expert in cement, but I don't think anything would happen negative, especially if the humidity is fine and sunlight is not getting in. The bottom bag of cement can lie around forever, so there, they would use a LIFO stacking system, where they would pick up the bag of cement that was put on last, because otherwise, it would be inefficient. They're heavy bags and to turn this whole thing over and create a queue out of cement bags might be an inefficient use of space.

Another one is plates requiring washing at a restaurant. So if you stack up plates requiring washing, you're obviously gonna wash the one that you put on last. You're gonna wash it first, LIFO. Another one is plates being loaded into a plate dispenser. You know those plate dispensers, those balancey things that we all like to play around with at restaurants or at buffet tables? So the plates that was put in last will be picked out first.

An in tray for an office worker. So if somebody's working at an office and they still have an in tray, instead of an email ... I don't think those are popular any more, but nevertheless, if you have an in tray, and people put items into your in tray, you're gonna pick up the one that was put in last, first.

And you can really see, as you're getting a sense for these examples, industry examples, you can really see. This one, for example, is specifically valuable, because if you're seeing inefficiencies in data processing or some process in a business, it could be related to somebody's in tray, just because they're getting overcrowded with items and they're picking up the last ones first, and therefore, the ones that should have gone first, or the ones that were given to them earlier, are being delayed. So there you go. There's an inefficiency right there.

Another one is you packing your travel bag. When you're going to travel, you pack your bag. Think about it. The items that you put at the top are the items that went into the bag last, but they're the ones that are gonna come out first.

So typically, when the entrance and the exit is the same, then you're gonna have a LIFO method. When there's an entrance and a separate exit, then you got a FIFO. That's one way to think about it. Another way to think about it is queue versus stack. There's lots of ways to think about it.

And so we just rambled through probably 10 different industry examples, and you might think to yourself, "Well, some of them are useful. Some of them, you can't really do anything about. That's just the way things work, perishable items this way, cement this way or that way." And in many cases, it's true. In many cases, it's just the natural way of things.

But what I'm suggesting here is if you get into the habit of picking up on these things throughout the day, throughout the course of your week, and just seeing, "Okay, I'm standing waiting for my burger or for my food. Is that FIFO or LIFO? I'm picking up this plate. I'm packing my bag. What is FIFO? What is LIFO?"

Once you get into the habit of doing these things in our day to day world, in these trivial cases, you will get a sense for it, how to do it in your business, how to quickly pick up on these things, how to even notice them. A lot of data scientists don't pay attention to this, and they go through understanding a business and a business process, and they don't really notice, "Is this FIFO or is this LIFO?" And they don't see the implications that that has. And the opportunities that that might yield, or the consequences it has on the way data is coming in and out of these processes, or the way that data is being handled, the way that data is actually being generated and collected.

And so I think it's important to at least be aware of these, and ideally, pay attention to them, and therefore get value out of these concepts and understand your industry, your domain knowledge quicker, and grasp those little details that are intrinsic characteristics, that other data scientists might have not picked up on, or you wouldn't have otherwise picked up on.

So on that note, I hope you enjoyed this podcast and this little excursion into the nerdy world of data science, or I would say even domain knowledge in data science, and hopefully, this was a helpful tip for you. And if it was, my challenge for you this weekend would be to try and observe your day to day and see where you can pick up on FIFO and LIFO and see how many examples you can pick up in just one weekend.

And hopefully, that will help train your mind to continue doing that next week, and then throughout the month. And just picking up on these little details and then seeing what that means for your understanding of your business and how that can enhance it.
On that note, thank you so much for being here today, and I look forward to seeing 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?

EMPOWER YOUR CAREER WITH SUPERDATASCIENCE

CLAIM YOUR TRIAL MEMBERSHIP NOW
as seen on: