Hobbs: Hello everyone, this is Hobbs from Valorem Reply. This is TNT and I’m excited for one of the first times to have a guest speaker. This is Robert Goss and he’s going to tell you a little bit about himself in just a minute.
For those of you who have not been here before, this is a video series where we talk about best practices and common mistakes in the data, IT and business intelligence space. So, with that, let’s go ahead and get started.
Welcome back everyone, as I said this is Robert Goss. Robert why don’t you take a minute and just tell the audience something about yourself, who you are, what you do, what your background is.
Robert: Yes, thank you Hobbs. I am Robert Goss. I am a Data Engineer with YRC W here in Kansas City, it’s a large trucking logistics company. I’ve been in the data space a couple years now. Before that I was in finance. I’ve been part of the data team at YRC W for about 18 months and working very closely with others on the team for business intelligence, some visualizations as well as doing some data science. We’ve deployed quite a few algorithms so far and been really excited about it.
Hobbs: Cool. Tell me some about your technical background. What are the programming languages you work in? What is your corner of the market if you will?
Robert: Yeah, so I would definitely call myself a rock star when it comes to Python, that’s my forte. I’m also big into SQL server and other relationship databases as well. I started out doing some web pieces as well then started exploring other data structures about a year and a half ago and have been pretty interested in that
Hobbs: Well Robert, as you know, this video series is all about common mistakes that us professionals see in the industry across clients and projects and what do you replace those mistakes with. What’s the thing that fixes that problem? So if you don’t mind, let’s start with what you see as a common problem in your interactions with clients and the industry.
Robert: Sure. One of the things I see a lot is the idea of best practices trumping an ideal best practice or ideal situational choice.
Hobbs: So tell me what you mean by that. Typically I would say follow the best practices right? So when you’re saying that a best practice shouldn’t always trump the situation, can you give me sort of an example of what that looks like?
Robert: Absolutely, so you may have a best practice to always use .net or SQL server. Today I wanted to really focus on data structures, so databases in particular. Sometimes people choose a relational database when there’s an alternative data store that is much better suited for the situation.
Hobbs: And I suppose if you’re always just going for the best practice, a lot of times there’s going to be a heavy cost invested in moving that direction, when maybe your best bet, do you think, is just to stay where you’re at and use what works for the business? What are your thoughts there?
Robert: Yeah so sometimes it’s good to actually investigate other ways to do things. So maybe your best practice is to use SQL server, but you may want to look into something like a no SQL database. Or there’s in-memory data storage for a shorter period of time. Graph databases, which are coming up in the world. There are even hybrid approaches where you could have SQL and no SQL in the same data structure.
Hobbs: So if you were working with someone in the audience that has started a new project and they’re thinking: ‘ok I’ve been told by all these people: here’s the best practice, here’s the way you should do things.’ How would you advise them to move forward comparing that best practice with their particular situation? What’s your advice for them?
Robert: I would focus on the business problem, understand where the business problem lies. If you’re dealing with an accounting team that deals in spreadsheets, you should probably understand that that’s going to be a relational data store, using a relational database is ideal. If you’re dealing with a network of sorts or you’re dealing with a social media manager, all these kinds of interconnected relationships are more ideal for a graph database or a no SQL database. You really need to understand where the business is coming from, the problem, and then look at structuring your data accordingly.
Hobbs: Ok, so if I was to sort of summarize that- because we always try to get things down to a short statement, this thing not that thing approach- would say that you’re saying that the business case or the business scenario is going to trump the hypothetical or theoretical best practices?
Robert: Yes absolutely
Hobbs: Alright, well I like that. For those of you out there that are starting projects, sometimes you’re going to be able to follow all of the advice that you get from people. But, sometimes your situation is different enough that you just have to do what works in your specific case.
Thank you very much for joining us today Robert, it’s good to see you.
Robert: Glad to be here.
Hobbs: And we’re going to get some more people on the video series as we move forward.