Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
015 - Chaos to Clarity: Navigating Data Modelling Challenges
314 Plays1 year ago
๐๏ธ In this week's episode of @the stacked data podcast, we dive deep into the world of data modelling and uncover strategies for success with our special guest Nieves Gorriti the Head of Data and Insights The Dot Collective
Here's a sneak peek into the insightful discussion:
๐ Introduction to Data Modelling:
What exactly is data modelling, and why is it crucial for effective data management? Nives provides expert insights into this fundamental aspect of the data lifecycle.
๐ Common Pitfalls and Challenges:
Discover the common pitfalls organisations face in the data modelling process and learn how to avoid them to ensure smooth data operations.
๐ Importance of Data Modelling:
Uncover the pivotal role of data modelling in driving successful data-driven initiatives within organisations, along with real-life examples demonstrating its impact.
๐ง Guiding Through the Process:
Nives shares invaluable advice and practical tips for enhancing your data modelling skills and provides a step-by-step guide to approaching data modelling successfully.
๐ Balancing Flexibility and Standardization:
Learn how to strike the right balance between flexibility and standardisation in data modelling to meet business demands while maintaining data integrity.
๐ค Adapting to Evolving Data Environments:
Explore strategies for adapting data modelling strategies to accommodate the ever-evolving data landscape, including new technologies and data sources.
Don't miss out on this conversation! Tune in now to Stacked Data Podcast to gain valuable insights into mastering data modelling for organisational success. ๐ง
Recommended
Transcript
Introduction to Stacked Podcast
00:00:02
Speaker
Hello and welcome to the Stacked podcast brought to you by Cognify, the recruitment partner for modern data teams hosted by me, Harry Golop. Stacked with incredible content from the most influential and successful data teams, interviewing industry experts who share their invaluable journeys, groundbreaking projects, and most importantly, their key learnings. So get ready to join us as we uncover the dynamic world of modern data.
00:00:34
Speaker
Hello, everyone. Welcome to another episode of the Stacked Data Podcast.
Interview with Nieves on Data Modeling
00:00:39
Speaker
This week, I'm joined by Nieves, the Head of Data at The Dot Collective. Nieves is going to talk about data modeling strategy for success. I know it's something that you're extremely passionate about and keen to dive into, because I think it's an issue and an important talking point for the data industry in general. Nieves, it's great to have you on. How are you doing? Hello, Harry. I'm very happy to be here. Good, good. Thank you, yourself.
00:01:03
Speaker
Good. Yes. Yeah. Very good and excited to dive into this day's modeling chats. One of my favorite topic areas. Also mine. That's good. Well, it'd be great for the audience if you could give us just a brief overview of your career and the Doc Collective and what you do at the Doc Collective and your role.
00:01:23
Speaker
Yes, of course. Well, I studied computer science, so already at Uni I started getting this love for data modeling. In fact, I never liked programming, so data modeling was really the thing that I really liked. And then I've done a lot of different roles. I was a sub-engineer for a couple of years, and I jumped into the analysis, business analysis, process analysis.
00:01:46
Speaker
And yeah, sometime back I came to London and here I got the opportunity to work as a data analyst. And yeah, of course the part that I enjoy the most was data modeling. And right now I'm working for the dot collective.
About The Dot Collective
00:02:00
Speaker
So the dot collective is a data consultancy company. So we are focused only on data. We are doing many different types of projects right now, but all of them are linked to data engineering, data analysis, data architecture,
00:02:12
Speaker
cloud engineering. So we are trying to deliver solutions for the clients in many different tools. Let's say we are not only focused on one type of cloud or one type of solution. Data migration is a bit of everything where we can help the customers. So that's what we are currently doing at The Dot Collective. We are growing a lot. I'm the head of data analysis. And as an example, we were eight.
00:02:34
Speaker
last year, and we are 17 right now. We are really growing very, very fast. It's very good. It's a challenge, of course, because data modeling, it's the most demanding thing from the clients, and it's not that easy to find data modelers. Let's talk about the reasons. Yes. I suppose if anyone's listening and is a keen data modeler and likes what they hear from the MS, then do reach out. You should be happy to talk.
Understanding Data Modeling
00:03:02
Speaker
First off,
00:03:04
Speaker
What actually is data modeling? Can you give sort of an explanation as to what data modeling is?
00:03:10
Speaker
Yeah, of course. So data modeling is pretty simple. It's understanding or giving an instructor to your data, making your data understandable. It can have many different ways of modeling it. It depends on the type of data. It can be structured data, non-structured data, images, emails, or simply CSV files, for example. But the fact of saying, OK, I'm storing this type of information here and this type of information here so I can use it later, yes, that's data modeling.
00:03:39
Speaker
There are many techniques. Well, many techniques. Finally, everything is quite the same, let's say. But it's just a fact of making the data available and clear, knowing where is the data. So making sense from raw data, from mess, making it into a condition where it can actually be actioned upon.
00:03:57
Speaker
Indeed. So for me, even raw data is modeled somehow because you have some structure. You are storing the data, for example, following days. But there are many ways to keep your data structured. So from raw data, which is just data from the source in a specific structure, to a very clear defined model where you say, these are my customers and these are my sales.
00:04:19
Speaker
everything in between, it's a way of modeling the data. So it depends, of course, on who's going to use it, how they're going to use it, how much data do we have. But the fact of knowing where is your data in any type of structure is just data modeling.
00:04:35
Speaker
Yeah, okay, that's brilliant. And you've already alluded to some of the points that we can cover.
Role of Data Modeling in Management
00:04:40
Speaker
But in your view, why is data modeling such a crucial part of the data management process and data refinement in general?
00:04:51
Speaker
Yeah. So for example, when I started my career, there was no IT department in the demo person. And basically they were just using Excel, which is okay. You can use Excel, but they had so much data that they were not able to record any new row in the Excel file. So nobody can use the data. It was very difficult to match their data.
00:05:14
Speaker
And the problem is always the same. Everybody thinks like, no, we are just going to have a very small amount of data. We don't need something expensive. But you have free databases. I mean, you don't really need to spend a lot of money. So initially, I think the problem is that initially, everybody thinks that they don't have such a big amount of data. So let's just build something very small. Let's just use Excel or something very basic. And then at some point, they keep adding things. New people, new teams wants to use data somehow. And then it all becomes like a big,
00:05:44
Speaker
Nobody knows how to access the data, who is using it. And yeah, we really need to understand that even if it's a very small amount of data, we need to put some management behind. We need to put some models. We need to define this is
00:06:00
Speaker
I don't know, for example, as I mentioned before, these are your customers. You are storing them in here. These are your sales. Don't mix everything, because at some point, things will grow. The data will always grow. We have data constantly. We have new data every second. So just keep that in mind. And if you start with a good base, then it will be much easier to keep growing.
00:06:22
Speaker
That I couldn't agree more. It's so much easier to start with a solid foundation and set that from the get-go. You mentioned obviously at the beginning, it's hard to find great data modeling skills.
Common Mistakes in Data Modeling
00:06:36
Speaker
From your experience, where do individuals typically go wrong in data modeling and what are the consequences?
00:06:44
Speaker
As per my experience, I've been working with many type of people with very different backgrounds. So for example, I did computer science, but most of my colleagues, they didn't. So they don't have this base, let's say. And they started learning data from other sources. And now what you will always find is a lot of data engineering, a lot of super fancy tools where modeling is not needed. You just put your data here, we will do your work. And then you can just start reading your pipelines or
00:07:14
Speaker
building your reports. And for me, that's the problem, because there are so many tools right now that build the model, which are very useful, but not for all the solutions. So right now, data modeling was something that was like the ugly friend. Nobody wants to be with data modeling, because why do we need to learn how to do data modeling if we have tools that do that?
00:07:37
Speaker
So we really find very good data modelers that are a bit older, like me, let's say, but people in their 20s, it's very hard to find good data modelers because they've been using tools that were doing the modeling for them. The big problem we are having right now, because we do have very good data modelers in the market, but they are less and less. So we need to
00:08:00
Speaker
try to promote data milking again. So this is why I'm also very happy that this podcast is happening because for me, data milking is my passion. I love it. And I think it's very important for every company that really wants to use their data correctly.
00:08:14
Speaker
Yeah, I mean, your data model is how you make sense from your data to actually answer business questions. So it's such a crucial part of that data lifecycle.
Impact of Data Modeling on Success
00:08:26
Speaker
You mentioned tooling. We've got the likes of DBT. It's taken the industry by storm, and I think it's a hugely, hugely powerful tool.
00:08:37
Speaker
I didn't want to mention the scenery, but while you also have the same, for example, in Power BI, it's quite easy to have everything there. The model is more or less built, so there are so many tools that right now build the model. But you said something very key. You said that the data model represents what the business wants.
00:08:55
Speaker
And this is something that most of the people forget also. I've worked with amazing data models that they built models that were able to support historical data, like very fancy models, but they never took to business. And what happened finally, that we needed to start from scratch because business couldn't use the model for what they wanted.
00:09:14
Speaker
So for me, key skills of a data modeler, first, talk to business, understand what they want. For example, two companies doing exactly the same. Their models can be very different. Their data models can be very different because maybe for one business, they want to focus more on data science and others, they want to focus more on analytical purposes. So the models will change, but maybe for one
00:09:36
Speaker
a company, the granularity of the data is different. So it's not just saying, okay, these are my sales, these are my customers. We need to understand what business wants to do with that, because that will be the base of your mother. So again, what's the problem of data management? There is a big differentiation or a silo between business and the technical teams. They don't talk to each other. I know what they want.
00:09:59
Speaker
How do you know what they want if you don't talk to business? So that's also key for me in data modeling. You can be a very good data modeler, but talk to your business. Each company is different.
00:10:09
Speaker
Yeah, that's key. Understanding the business question is essentially how you're going to drive value. How does effective data modeling contribute to the overall success of a data-driven initiative within an organization? We've alluded to bridging that gap, but yeah, I suppose we're keen to dive a bit more into that.
00:10:33
Speaker
Yeah, of course. So when you have a good base on your data, you can build many new opportunities. You can build data products that are valuable for your business. There are thousands of data products. Data products are basically something that consumes, transforms, and generates data. But a company wants data products that provide some return on investment.
00:10:54
Speaker
So if you have a lot of data but you don't know how to use it, you don't specify who is using it or any security behind, nobody is going to use it. Nobody wants to use it. However, if you have a clear base and a clear documentation explaining where is the data coming from, who can use it, how can they use it, they will have more and more ideas. Business will be able to play around with the data and say, oh, we can't do this or we can't do that.
00:11:17
Speaker
So for example, in another past experience, I've been working for a highways company, and basically they have information every second about what's going on in the roads, if there is an accident, if it's raining. So there is so much information going on in there. Can you imagine the amount of things that you can do with that data? You can prevent accidents. You can, for example, if there is some security, transportation needed, they can organize it securely. They can have the Royal Mail with traffic jams.
00:11:45
Speaker
There is so much things that you can do with a good base of data. But you first need to understand, OK, what's my data? What can I do with it? How much data do I have, which is also key? Do I have data every second or every day or every month? And then with that, you will be able to power up your company.
00:12:04
Speaker
So it's essentially, it's so important to give you that agility in terms of how you want to answer business questions. If you don't have a data model that gives you that ability to almost pivot in what's needed and answer different questions, then you're obviously stuck in a very sort of rigid format.
00:12:23
Speaker
Yes, start with a very basic use case, because you cannot move all the data or prepare all the data for a company in one month. So you start with a basic use case, but a use case that is valid is very valuable for business, it's worthy. And then...
00:12:37
Speaker
When you allow them playing around with the data with an easy tool, with an easy report, with something very basic, they will start having ideas, oh, I can also do that. It's not just a technical team saying, you can do this with your data. It should come from business. But we need to explain to them what can be done. So we need to give them the tools also to play with the data. And the data needs to be easy to understand and flat and accessible.
00:13:00
Speaker
How do you and especially your position in a consultancy, how do you translate the value of and the importance of data modeling to maybe an end client, a business stakeholder, someone who's not necessarily from even a data background and you have to explain, data modeling does take time, takes time to get that.
Communicating Data Modeling to Stakeholders
00:13:20
Speaker
How do you explain that value to stakeholders and why you need to follow this process?
00:13:26
Speaker
Yeah, that's very tricky. And in a lot of conversations with new clients is why you are taking so much time in this initial analysis phase, while other consultancy companies, they don't need so much time. Because we are trying to build a good base. So the first thing we try to understand when we go to a new client is, OK, what are your current pain points?
00:13:49
Speaker
So we spend a lot of time with the client running workshops to understand the pain points, to understand what will be the ideal work scenario. What do they want to do with the data? And they currently have. There are dependencies. So we spend time with them. And also something for me very, very important in this initial phase is to understand if every team calls every entity the same way.
00:14:12
Speaker
Because, for example, if the sales teams, they call the products or the clients' clients, and I don't know, the reporting team call them customers, you are calling the same entity in two different ways. So we also spend time explaining the importance of setting a taxonomy or ontology model, explaining what's the base for the whole company, making the whole company working together for the data. Instead of just arriving and saying, hey, this is your data platform, here's your data, no, we really want to
00:14:39
Speaker
deliver a solution for the client, a solution for the whole company. So we spend a lot of time talking to everybody involved in the data, so everybody in the company, to ensure that the definition, the terminology is clear, and they know what they want, let's say.
00:14:55
Speaker
That's good. So it's a continuity. That's what I suppose a core data model provides a business, that continuity. We're all talking about the same, we're all on the same page and we're all answering the same questions. Great. Okay. So we've already mentioned that some of the pitfalls and some of the reasons with the issues in the industry at the moment are
00:15:15
Speaker
from tools like DBT, which have lowered the bar to entry, which is great in some ways, but I think is also missing out on some of the core components of actually what data modeling is. For those looking to enhance their data modeling skills, what advice would you give them in terms of where to start and what to focus on?
00:15:36
Speaker
Yeah, that's a good question. And I know there are very different profiles, so I feel it in my team. When we hire people here, there are some people that are very good, for example, with SQL and Python, and not that good in data modeling. So I always try to help them, like bringing exercises or using some YouTube videos. There are a lot of YouTube videos available with very basic data modeling.
00:16:00
Speaker
And some people are like, I prefer to watch videos. And other people are like, I prefer to do some exercises. So for everybody, I would say, let's start with some very basic videos in YouTube. There are very basic ones about entity relationships. What is third normalization from? It's just 15 minutes video with a real example where you see, ah, that's the point. But then I really, really encourage everybody to download a data set from Kaggle from many open sources. You can download a data set.
00:16:28
Speaker
which is normally a CSV file, and then check and say, how can I structure this? What are we having here? So start, what are the main entities in my data? And then how can I build the database and just build the database? All the information is in the internet. Everything is there.
00:16:44
Speaker
So my key advice is always download that data set and try to build the model. And then we'll discuss how we can enhance it. And then, of course, you have evolutions. Then you can start looking for slowly changing dimensions, scheme value, different ways of modeling. But let's start with a very basic entity relationship model, and then we'll keep upgrading your knowledge.
00:17:06
Speaker
Amazing. I was going to say, how important is it for people to know specific philosophies like sort of in the dimensional modelling? There's many different approaches out there. It's something that we often find you see data modellers don't even, some people haven't even heard of within the industry, especially the more younger generation. So yeah, how important is it to know specific frameworks and philosophies about data modelling?
00:17:31
Speaker
I think it's a bit like everything. People will fall into them logically, because they are quite logical. So depending on what you want to do, you will fall under Kimbal or email models. For me, those are the basic ones, and then you can evolve a bit. So if you want to do analytical platforms, and you want to read your data very fast, you want to have only one table. The less joins, the better.
00:17:53
Speaker
So some people will automatically test it with a lot of data, and then they will see that the performance is better if they put everything on the table. Maybe behind, they don't know that they are doing one way of modeling all the other.
00:18:06
Speaker
My advice, which I think it's something that right now we have the access to everything in the internet and people are just watching super things. It's important to know them, but they are quite logical. At some point,
00:18:24
Speaker
it's important to understand the difference between them when to use one or the other, but I think as they are quite evident. That's an interesting one, especially these people that are maybe a bit more self-talk. They often are doing maybe the right things, but they don't know then how to talk about what they're doing. And that's what lets them down in an interview I've seen in cases. They know what they're doing, but they don't know how to articulate that in a technical or non-technical
00:18:51
Speaker
way.
Strategy for Successful Data Modeling
00:18:52
Speaker
So yeah, maybe, you know, as you said, jump on the internet, make sure you understand, you know, the official approaches of what you're doing is then you'll be able to articulate it. So could you outline a step by step guide or, you know, best practices for approaching data modeling successfully within a projects and, you know, what's your strategy at the dot collective for this?
00:19:13
Speaker
As we discussed before, we first started by understanding the business. What is the business? So for that, you discuss with business, you also search a bit in Google, so you try to understand what are they doing and who are they users and who is using the data. So you organize works with the team, you understand the pain points, you understand what they want to do with the data.
00:19:33
Speaker
You understand if everybody's aligned in the terminology. Do they need a taxonomy or not? And then based on that, my first reflect is always to build a conceptual data model. So the concepts of the company, the main entities, I have my products, my customers, my sales, I relate them together. I don't need so much granularity like customer type in that conceptual data model. It's just how the company works. It's something for everybody in the company. You don't need to be technical to understand a conceptual data model.
00:20:01
Speaker
Then to understand what business wants to do with the data. If we are talking about data science and MLOps, let's focus on a data lake maybe, because you can drop so much data in their structure, and some structural data in there that can be useful for their data science activities. So let's maybe focus on a data lake. If they want to build an analytical platform,
00:20:21
Speaker
Let's focus on a data warehouse. Let's focus on data maps. How do we build them? Do we need different stages? Right now, normally, we always deliver in three different stages. So you have proven silver and gold where proven is your raw data. So in case failure, you can always come back to your source data. Then you do some transformations. You prepare your data in your silver area that can also be used as a data lake somehow.
00:20:44
Speaker
And then you have a goal, which is your data warehouse. You have all the data there with all your history changes. So you can understand what was the status of my company yesterday and today.
00:20:55
Speaker
And then, based on that, let's discuss again with business. Okay, you have your data here already. You can do all this. What do you want to do next? You want to do some reporting? Okay, what's the information? A data warehouse? I always compare a data lake. I like that in which I always have it in my head. A data lake is like this type of drawer that you have in your kitchen and you put a lot of things that you don't really care about the hardware. You just have your stuff drawer.
00:21:20
Speaker
In the lake, let's say it's very easy to drop things in there, but then if you want to find something there, you will need some time. But that's okay because it's for data science. So you don't really mind spending time with them, but you just want to write data very easy. A data warehouse is these big one drops that I think everybody would like to have some time.
00:21:40
Speaker
It's a walk-in ladder where you've got everything organised, you're all labelled up. It's full of things. You have a lot of things, too many things, but you have everything you need. And then maybe you need a data mart, which is for this report, I only need this area of my wardrobe. I only need my sports clothes to do sports.
00:22:02
Speaker
So that will be your data mart. And basically, you will need, again, that's for business. You will need to understand from business how do you want to use the data from all these data warehouse, which part do you need, when do you want to access, and then we'll build this small data mart. So this is, let's say, the process. We start with business.
00:22:20
Speaker
understanding the business, the taxonomy, we serve with a conceptual data model, then we understand how they want to use the data, we build the models. But the last bit, which is very important, is we run UIT with the business. Because, yeah, the data is there. And how do we know that we understood correctly? Because
00:22:39
Speaker
Finally, consultancy companies, you are with the client during a limited period of time. So you try to understand what they need, but we are all humans. We can all make mistakes and understand messages in the wrong way. So I always encourage all my... Well, encourage, no, oblige. But the title is to run UID. So build a set of test cases, build a couple of reports, depending on the scenario, and ask business to play with the data before going into production.
00:23:06
Speaker
to ensure that everything is good. Because otherwise, what we are building is not useful. And this is the moment where you will really discover if your data is performant, is secured, fulfill what business needs to do, is understandable, is accessible. And then they will start having so many use cases, so many new ideas that we really enjoy playing with the data.
00:23:26
Speaker
Amazing. I mean, it's clear that the ability to communicate with the business, understand their problems, and then understand, you know, from there, you can then develop what solution you actually need and is relevant. So that's become clear and spend a lot of time understanding what you're actually trying to achieve. In the building trade, it's, you know, you measure twice, cut once. So it's the same with data modeling, right?
00:23:49
Speaker
Yeah, you need to test it, because otherwise you can understand something and learn business. But it takes two minutes to filter my reports, like, what do you need? Okay, so look, the MS data modelling often involves this cross-collaboration between various different teams.
00:24:09
Speaker
How can organizations foster effective communication and collaboration amongst their data modelers, analysts, and stakeholders? Because it's clear that this is probably one of the most important parts of getting it right. So how can you effectively facilitate that?
Collaboration in Data Teams
00:24:27
Speaker
Yeah, that's a very good point. And it's very complicated. I consider myself lucky. I've been in many different companies because I've been working in consultancy companies for a long time. I don't think I've found a company know what all the teams are doing in general. It's very difficult to make a company talk to each other. And I don't really understand why, because they should be using the same data. They should be focusing on the same goals of the company. But well, it's understandable, because big companies at some point, they grow so much.
00:24:57
Speaker
These things happen. You've got these silos. Yeah. For me, what is key, which is kind of what they are talking about data manage, et cetera, is that each team needs to understand their data, the data they own. Because I guess that in a company, you will have sign-on, same human resources, many different things. But they produce their data, or they use their data. So they need to understand what data they use, what data they produce, and if it can be useful for other teams.
00:25:23
Speaker
So at some point, I think that the role of a product owner or product manager needs team that talks to other people. So that person for me is the key one understanding what my team wants to do and what my data team is. My data of my team is. Because then they will be able to talk to other teams and say, hey, we can share this. We have this information. Don't do this. We already have a report about that.
00:25:50
Speaker
So for me, there should be like key people in each department that knows their data, knows the requirements of the company and are able to share it with the other product teams. Perfect. It's very easy to say.
00:26:01
Speaker
Yeah, easy to say, hard to execute. But I think it's something that as a data modeler or as a leader, whoever is listening to this, it's just food for thought. How can we foster that and increase that? And what can you do personally to the work that you're doing
00:26:21
Speaker
How can you communicate more and understand more? And I think it always pays dividends to understand what other teams are doing. It helps you with your own understanding of what the overall mission of the company is. But there are some processes that can be also implemented in a company. So I'm not a fan of large volumes of documentation because it gets out of date very, very quick. But at least you need a minimum. Build a logical data model.
00:26:47
Speaker
Ensure that the conceptual data model is at the company level. That should be the first thing. But then try to build a logical data model and keep it updated. Or try to use tools that reflect the current mappings of your databases. So ensure that at least you have a minimum documentation that reflects the current status of your database and your data.
00:27:10
Speaker
And then also, when the companies are implementing these processes, you can keep evolving. And you can start, for example, building data contracts. So this is the data that I have. This is the people you meet. This is the security that I need behind. These are the SLAs, et cetera. So then you will have data contracts at the company level. But simple ones. You don't need 50,000 pages in a data contract. Very simple ones that can be updated easily, that are shared for the whole company.
00:27:38
Speaker
So at least a minimum documentation that will help other teams understand what's in there. Yeah, simplicity, I think, is the key point there because you can overcomplicate these things and you pull up a document which is x amount pages long. It's not going to be used again. Indeed, indeed. And then you need to update something. It's like, oh, how many times did I write it down in this document? It doesn't make any sense. Yeah.
00:28:02
Speaker
Perfect. Well, look, final question there, Sim. The ever-evolving data landscape is constantly changing new tooling, new approaches. How should organizations adapt their data modeling strategies to accommodate for any new technologies, data sources, changes in mission? Yeah, what advice have you got for that?
Adapting to New Technologies
00:28:22
Speaker
Yeah, I think we started the conversation saying that there are some tools that because of those tools, data modeling is dying. But I don't want anybody to think that data models are against new tools. This is not the case. So the tools are there to help. And you can use new tools
00:28:40
Speaker
to, for example, improve this documentation. You have a lot of tools right now that can define your data image in a company. You don't need to write it down in an Excel file. You have a tool that will build your data image, that will build your security, who has access to its data. So there are many tools, and there are more and more, that will help you keeping some data governance in your data automatically. So I'm always very open to new tools, and I'm always trying to play around with new tools because
00:29:08
Speaker
I never found the perfect tool that is able to do everything. But little by little, there are new tools that will help you having this documentation, sharing the data, sharing the data, contracts, sharing what is in your data. So yeah, it's always a matter of keep trying, keep testing new tools. Some of them will be good for something, other for other things. So it's not, I just want to keep my data model in my old ways of doing it. No, always try to use new tools, because they will be very helpful. Building data models.
00:29:37
Speaker
a diagram is easy. The problem is always when you want to do reverse engineering to understand what's the existing data model and what are the transformations because they are not documented, or when you want to improve your model or build all the mapping files. So if you have tools that do that for you, use them, please. Yeah.
00:29:55
Speaker
Perfect. Well, let's try and stop data modeling for them from dying. Yeah, so you've given some really great advices to really what data modeling is and why it's so important and how people can level up their game. I think it's a valuable skill to have and is so important in the modern data world and for driving value. So thank you for your insight.
Career Advice for Data Professionals
00:30:18
Speaker
We now move to the quickfire round. So this is a section we ask all of our guests. The first question here is, how do you assess a job opportunity in your career and how do you know it's the right move for you? Well, that's really difficult. You will never know if it's a good one.
00:30:34
Speaker
Even if you ask a lot of questions in an interview and you want to understand, this is what didn't work in my previous work, and I want to know if it will work in this one, you will miss a lot of questions. So you will always have a risk. But if you feel like they will train you, they will spend time and resources helping you progressing in your career, for me, that's a must, and that's a very good opportunity.
00:30:58
Speaker
The second question, Nia, this is one that I think a lot of listeners really enjoy understanding. What's your best piece of advice for people in an interview? Be themselves. I'm always getting very nervous in interviews, but just be yourself. You know the answers. And if they're asking questions from a book,
00:31:17
Speaker
then it's not your place. Why do you need to know what's the meaning of, I don't know, what's the second normalization form exactly? You don't need that. You just need to be yourself and demonstrate that you are able to think and say, I don't know now. I will check in Google when I need to. And that's all. For me, that will be my advice always. Try to be yourself. And if you don't know something, it's OK. Nobody knows everything. But just be yourself and say it.
00:31:42
Speaker
Yeah, I think that's great. I mean, the being yourself is always easier said than done with nerves, but showing your ability that you can upskill. And I think giving examples of where you've been able to do that is a natural way of them when you don't answer or you don't know the answers to the question. Give an example of when you haven't known an answer and you have been able to quickly answer it. Finally, if you could recommend one resource to the audience to help them upskill, what would it be?
00:32:10
Speaker
Well, as I said before, I always prefer to find exercises and put your hands on. So download data sets and start playing around. But of course you need a very basic base. So you have very good books, but I will really recommend YouTube videos. There are a lot of people that in 15 minutes you have a very good video to start with.
00:32:33
Speaker
Yeah, I think I'm a visual and visual and audio learner as well. So, you know, people that don't necessarily like diving into books. It's a really quick way. It's something you can listen to as well without necessarily having to watch everything. So amazing to have you on the MSN. Your insights have been great. And yeah, it's been a pleasure speaking about an area we're clearly both passionate about. So thank you for your time. Thank you very much. It was really nice talking to you. Bye everyone. See you next week. Bye.
00:33:03
Speaker
Well, that's it for this week. Thank you so, so much for tuning in. I really hope you've learned something. I know I have. The Stack podcast aims to share real journeys and lessons that empower you and the entire community. Together, we aim to unlock new perspectives and overcome challenges in the ever evolving landscape of modern data.
00:33:24
Speaker
Today's episode was brought to you by Cognify, the recruitment partner for modern data teams. If you've enjoyed today's episode, hit that follow button to stay updated with our latest releases. More importantly, if you believe this episode could benefit someone you know, please share it with them. We're always on the lookout for new guests who have inspiring stories and valuable lessons to share with our community.
00:33:46
Speaker
If you or someone you know fits that pill, please don't hesitate to reach out. I've been Harry Gollop from Cognify, your host and guide on this data-driven journey. Until next time, over and out.