#127 How Data Scientists Can Thrive in Consulting

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com. You're listening to Data Framed, a podcast by Data Camp. In this show, you'll hear all the latest trends and insights in data science. Whether you're just getting started in your data career or you're a data leader looking to scale data driven decisions in your organization. Join us for in depth discussions discussions with data and analytics leaders at the forefront of the data revolution. Let's dive right in.

Welcome to Data Framed. This is Richie. A frequent theme of this podcast is to try and encourage you all to gain some data skills so you can solve data problems yourself without having to rely on others. Today we're talking about the opposite thing, using data science at consultancies where the whole point is to solve other people's data problems. In a previous stage of my career, I briefly worked as a data science consultant and I found that there are a lot of interesting challenges that apply to this field, like having to build trust with new clients and juggling multiple projects under tight deadlines. Since successful consultancies necessarily have to become great at data communication and data project management, there are lots of lessons to be learned for other industries too. My guest today is Pratik Agrawal, a partner at Kearney Analytics. He's a software engineer turned data scientist and has led teams of data science consultants at Boston Consulting Group. And now Kearney. He's an expert in digital transformation programs and digital strategy. So I'm hoping to pick his brains on how data science works in consultancies. Hi there Prateek. Thank you for joining us. Just to begin with, can you tell me a little bit about what Kearney does?

Kearney is a global management consulting firm. We are over 5,000 folks in more than 40 countries. Sizable presence in quite a few geographies.

Brilliant. So can you tell me a little bit about what you do as a partner at Kearney?

Absolutely, Richie. I as a partner at Carney and specifically in the Carney analytics practice, lead our automotive and industrials transportation sector and what that means is understanding our clients needs in those sectors as it pertains to a wide variety of topics, be it supply chain, be it manufacturing, as well as understanding the customer and the growth side of things as well. It requires us to understand where our clients are headed or even push sort of the thinking on where the clients can actually start exploring both new markets as well as gain new efficiencies or gain efficiencies and what they already do.

Can you maybe give us some details on examples of things that your customers are trying to solve in the automotive, transportation and industrial sector?

So if you think about, let's take manufacturing as an example, we're undergoing the fourth industrial revolution or in I 4.0 and a lot of different marketing terms. I'm sure either you've heard or the general audience has also heard. But what that entails is how are we helping optimize processes within manufacturing as AI is coming of age? How do we bring it to typically areas where analytics didn't really play a part? It wasn't necessarily infused. Digital was not a part of the discussion. How do you use things such as computer vision to understand safety violations or safety issues? Labor productivity, how do you understand bottlenecks in a plant? And can you do early detection? Can you build early warning systems? How do you organize your inventory within the plant to be able to say, I don't have to shut down my plant because I didn't have the right parts or the right labor available? It spans a bunch of different topics, but each of our clients typically tends to go through that journey, starting with one topic and so on, so forth. There are aspects of it that can be built into a program, such as Factory of the Future program, which basically entails analytics, continuous improvement, automation. And this is all on the manufacturing side.

So it sounds like there are a lot of optimization tasks there and just really preventing or having some warning about if something disastrous is going to happen. You want to know upfront before that disaster does occur?

Absolutely. I mean, it's predictive. It's also getting ahead of the curve is simulation. Can we simulate what our world will look like tomorrow or in a month's time? And if we can simulate, we can get ahead of planning, get ahead in our planning discussions. And that is a frontier that we're pushing with our clients.

Brilliant. I'm really curious to know how analytics from the consultancy side differs from when you're doing analytics for your own business. So perhaps can you talk me through at a high level, what does a consultancy data project involve in the industry.

What I saw as working in analytics teams, machine learning teams or AI teams was essentially that you had a singular focus. You weren't necessarily looking to go span across the enterprise. There were very few opportunities to essentially help optimize what one would do as a firm. And I think as a consulting firm we come in with a different perspective as let us look at the business overall and help you optimize that portion of it. And instead of going with just the strategic tools, let us bring in analytics as well. And that spans across the enterprise versus let's say having a more of a focused view on where analytics could apply.

That's really interesting I guess because you've seen lots of different businesses, you have this high level vision of what it can look like, is that correct?

Yeah, absolutely. And I feel that it's definitely been the unlock for my own personal journey. Having worked both in industry and then switching to consulting. What I saw that I could understand the value that I was bringing very quickly in these shorter projects, shorter programs, versus I think working towards a project that would be typically a year, multiple years if I were to be in industry. At least that was what my experience was.

And maybe just to make this more concrete for our listeners, can you give some examples of projects you've worked on, maybe something you've enjoyed or had a big success with?

A ton of projects that were all interesting, but chiefly among them. One experience that I always recall is my experience working in South America. This was for a large forestry client and an amazing opportunity to spend time on the ground working with the teams that were very diverse, coming from different countries, different backgrounds, different educational profiles. And what we were doing was helping this client go through the digital journey that I was mentioning before helping their plants enter the digital age. Where is essentially things like can we help you reduce production waste? Can we help you improve your quality? So essentially yield and all of that using real time AI, using the cloud based systems, developing our AI applications, but then pushing them and having them in cloud systems to do essentially round trip calculations in less than a second. That was eye opening both for the client and for us to see that this was definitely possible and it was driving value. At the end we were able to achieve a few percentage points in reduction and raw material required. Really interesting project. Loved the culture, love understanding who the people are and what the work culture is. Somewhat different than what I was used to as working in the US and that's what opened my eyes. It's like different geographies have different ways of approaching problems as well as how to come up with a solution and implement solutions.

That does sound like a fascinating project. I guess you're mixing the very high tech stuff with the AI and then you've got, I guess growing trees is a by default a sort of fairly low tech thing. It's trees. So mixing the two does seem like a very interesting project. I'd love to get into more detail about how you do that perhaps later, but before we get to that, I'd like to talk about who your team is and who gets involved in these analytics projects. So perhaps can you tell us a bit about your team? Who's working for you?

So usually these start out with as one would think consulting engagements would be accelerated, fast paced versions of what a typical sort of due diligence might look like. Or when you're assessing whether a project actually has or a problem has legs and has the right solution, can have the right impact. So starts off with a pretty small team of, I'd say like a manager plus one consultant or so and they'll be looking through both the business value of what we're trying to solve for actually even prior to that, what the problem is, defining the problem, putting sort of the bounding box on that, then going ahead with understanding what the value would be, starting to help prioritize what do we actually work on? Do we go ahead and work on the first problem that we encounter or do we go ahead and start prioritizing based on what's going to be the overall impact for our clients and from that prioritization matrix, essentially helping the client through on building those solutions With a somewhat larger team, I'd say still limited in terms of the number of folks that would work on an engagement, but somewhat larger team with which would constitute of data scientists, data engineers. Typically if you are in a, in a full blown scaling and implementation mode it's. It requires software engineers, UI experts usually you know, myself as an engineer and as a data scientist, I'm not really familiar with what works for users and in terms of interfaces and so bringing a UI expert is amazing because they can interact through interviews and start building out what is actually required for that particular problem. And so like bringing in the UI folks, the full stack developers, it's a massively joined team in terms of the skill sets that one needs to bring to bear. So it's not just the data science and the AI aspect of it, but it is all these different pieces that need to move together.

So this is actually really fascinating that you have this dynamic team that starts off with a few people and then grows to a much larger team. And I guess working at a company, that feels slightly unusual to me because I'm part of a team and that's like it's the same people all the time. So I'm curious as do people work with each other on more than one project? Then do you find that people tend to cluster and repeatedly work together or is it different teams every time?

I think that's an excellent question. I mean, the way the consulting model is set up, it is that if you're early on in your career, you get to experience a ton of different projects across multiple industries. And that's actually encouraged. You don't start aligning yourself that quickly to a said sort of industry practice or industry itself early on in your career. If you're just getting out of college, if you're just getting out of grad school, you're encouraged to spend more time across different engagements at a variety of them across functional, different functional topics. What that does is it helps you build a much broader base of understanding of what you can accomplish and what you can do as well as you start thinking about what you would want to specialize in given sort of what all you've seen. So it's only with tenure that you start specializing and you start seeing folks coming together on certain topics. So as an example, our automotive segment, if I take that there are partners who are already working in the automotive sector, have said clients, they decided this as either as an industry for them to focus on or this was something that they had experience, prior experience in as they stepped into consulting. But we continuously do work in automotive as partners. But then if you think about, let's say principals and managers, et cetera, you start slowly specializing into a specific industry.

It does sound like there's a sort of fairly well defined career path straight from being a graduate out of university to being someone who's a lot more senior. So that kind of career progression is very cool. I'm curious as to what the mix of people with data skills is and the other maybe less technical consultants.

One thing that I've found that has changed over these last many years in consulting is we've become more accustomed to working with data and this is just broadly speaking. So it's not just the analytics practices or the AI practitioners, but it's the broader sort of consulting landscape. And that's to do with, we think about where we can drive decisions from. Data driven decision making is, is, I think is baked in now in, in almost all parts of Consulting. Having said that, there are specialized skills that are required by analytics practitioners who are coming in because you're now the tip of the spear in terms of how you're both understanding the problem as well as defining what the solution might look like. And with that, thinking through on can you handle larger data sets? It's taking things beyond Excel. I think we were at Excel a few years or many years ago, I think it's taking it beyond that. And if you are somebody who loves doing that, I think there's definitely a huge, huge place for you to be in a consulting team.

It sounds like the difference between data analysts and the rest of the consultants is blurred a bit. So it isn't just people with data in their job title that need data skills. Actually everyone needs some level of data literacy. So can you tell me a bit more about what are the common data skills that people need?

Absolutely, Richie. I think, you know, the, the fact that we are approaching or we have already approached the point where there are advanced skills in data literacy that are needed or, or just advanced data munging skills, I'd say. And on that front, having somebody having skills in Python, somebody having skills in handling larger data sets that span more than just a million rows or so, are par for course. Now as an interesting example, now we've had clients who've asked us, can you handle a terabyte of data? Can you handle a petabyte of data? Or rather the request that you've made spans this much volume of data and trying to find that skillset to be able to handle that kind of volume is always, always in demand. So I'd say somebody who has skills on handling big data, has skills in good level of Python programming. That's the kind of data literacy that I typically look for who isn't fazed by the volume of data, who isn't phased by the challenges that come with data. That's what I typically look for in folks.

That's really interesting that Python skills and big data can just the entry level that you need in order to start working in this field, because that's quite a high requirement compared to some other fields. So related to this, are there any domain specific skills that are needed? So for example, do you need experience in automotive and manufacturing in order to work with you?

That's a great question. I think domain wise, somebody coming fresh out of college or some fresh out of university as a grad, you may not have the exposure to the domains that we typically have. Right. Like the different domains. But you know, maybe you've worked in a specific domain as an intern. But what happens is that as you're progressing, if you're coming in as an experienced hire, we're looking at you as somebody who has the experience in a set domain. So the domain knowledge can actually differ on. On two aspects of. One is more sort of generic domain knowledge is like what is the value chain for a. Let's take as an example automotive. Automotive has a very defined value chain on who the different players are, what OEMs are, what the different tier one suppliers are, how does the ecosystem work with dealerships, et cetera, Somebody coming in with that kind of knowledge, how does sales work? That, that would be fantastic, right? Or you could develop that knowledge while you start your journey with carney. The second and more technical aspect of it is there are different systems that are very specific to each of these domains. So having exposure to them is definitely a huge plus point. But not to say that you cannot gain that expertise again while becoming a consultant or being a consultant. I'll take as an example. Working in manufacturing, it's very common to be working on systems such as historians and historians are specific systems around recording sensor data, sensor signals at a very minute scale. So like at a very. At almost like a second level or even, even lower frequency than that.

So it sounds like the traditional flow then is that you learn the technical skills and data skills first and then you learn the domain skills after that. And I do find it interesting that particularly in manufacturing you've got this sensor data is incredibly important. So it strikes me as being a little bit like the Internet of things data that's becoming very popular at the moment.

Oh, that absolutely is right. It is the IoT or in some references it's also industrial Internet of things. So IIoT and we're quickly getting into now the Metaverse and how digital twins play a huge role in going beyond just utilizing sensor data.

I don't know anything at all about digital twins, so perhaps you can elaborate on that.

I can talk about digital twins in more sort of the manufacturing and the supply chain aspects of the world and what that entails is a digital representation of how the world around you works. Let's take a manufacturing plant as an instant understanding how each operation in your plant works, understanding what's the cycle time for every possible job or task within the plant and being able to codify it into a model which is essentially a clock driven or a system clock driven, let's say, procedure. You are now able to quickly step through the machinations of what a day might look like in the plant if you know what you need to produce, if who you have in terms of producing. So what's the supply as well as the folks and you know enough about your assets, so things like the machinery that are present in the plant and you know how they perform, you're able to then quickly go ahead and say the next shift or the next day is going to look like this. This is where I'll have bottlenecks. This is where I'll have an asset that's going to be down. And that is incredibly powerful because you can now on the fly, decide what you need to do in order for you to preemptively deal with the situation or de risk the situation.

Okay, so the idea is it's like a sort of a simulated version of your existing factory or perhaps part of it. This sounds like it might be quite useful for scenario planning.

Absolutely.

So do you use it for trying to understand different use cases like, okay, if our orders go up by 20%, then we can figure out how to deal with that? Is that the idea?

Absolutely. So if you've got what kind of mix of products you're going to be creating or manufacturing and you know who's going to be coming into the plant, you are able to go ahead and say, I can sufficiently do this or I'm going to be under and my throughput's going to be lower. And if you know that, then you're able to actively do scenario planning on like, let me think about shifting the demand to a different plant that might be able to meet the production demand and it won't suffer on throughput loss. Now that's, that's one use case, but there are multiple use cases that are of you can think about. For the digital twin, it's what are the opportunities to do improvement within the plant? What are the opportunities about? If I were to add another area, another sort of production facility or another packaging facility, where should I add that across my network? So now you're taking it beyond just the four walls of the plant, but you're thinking more on a network basis. So a ton of questions that can be answered by digital Twain things like, what should my new facility look like? So greenfield applications, if you've studied your existing setup, how do you optimize that?

That sounds like a really fascinating field. I'd like to pivot the conversation back to where we were talking about skills. So we talked a little bit about the data skills you need. So things like Python and working with large data sets. We talked A bit about domain knowledge and how that tends to be acquired over time. I'd like to talk a little bit about the soft skills. So are there any soft skills that you find to be particularly important as a consultant?

That's a great question and a great point that you're hitting on, Richie. Soft skills are paramount to. I think whatever we do being successful. A large part of consulting is change management is to help our clients embed whatever has been developed and ensure that there is sustainability of those solutions. So as far as that goes, soft skills, being able to communicate clearly, crisply, articulate what is required, as well as, I'd say communicating progress and really being there for the client and having an active ear for them is really important. A lot can be quickly accomplished if you've got the right sort of communication skills.

Okay, so it seems like communication is perhaps the number one soft skill that people need. Is there a difference between communicating within your company, between team members, and communicating to a client?

I'd say that within the team itself, communication is everything. Communication, being able to collaborate effectively, being able to communicate as quickly as possible, and not hesitating from raising your hand if there is something that seems that you might not have familiarity with. It's really important to be able to communicate as quickly as possible. And one shouldn't shy away from that because what that does is it immediately gets the team involved, it brings out a lot of discussion, and you have your answers very quickly as opposed to trying to solve things by yourself, which can tend to be a little bit longer since you're trying to research topics which you don't have. Familiarity. It's just basically crowdsourcing, I'd say, like crowdsourcing. If that's a pain point for you, you've not been able to be able to come up with the right solution as quickly as possible. Crowdsource, understand where. Where pockets of expertise might be and. And leverage them. And with clients, I think it is. It's also one important aspect is let's co design with clients so it's again, do not develop something in a silo because ultimately the client has the right domain expertise. Has the expertise right. They've been in the firm for a while, they know exactly how things work. Involving the client themselves in this design process, making them a part of that journey is also paramount to making these projects successful.

All right, brilliant. And just to wrap up the section on skills for anyone who's interested in working as a data consultant, what do they need to do to get hired by you?

Well, I think Three, three key things. One, be genuine and genuinely interested in the problems that you're trying to solve. So be there for the client. Be interested in what the problem is and helping them. The second is be curious. And by that I mean curious enough to explore the art of the possible. No need to shy away from innovation or pushing the envelope in terms of what can be done. Frequently we do get inspiration from other sectors and try to bring it to our own clients. So being curious is really important in trying to. And the third I'd say is collaboration. So being open to collaborating, that is, that will get you the most learnings as quickly as possible in your journey as a consultant. Now, in terms of data itself, I'd say have the technical skills, but these three points, I think these three points are really paramount to you becoming successful and getting hired as well.

So moving on to talking about how consultancy differs from other businesses. It does seem to be a fairly unique beast. So I'd like to know if there are any challenges to working with data that are specific to consultancies.

Well, I think the time pressure, Richie, there's always a timeline that one is working against. And so working with data in a consulting firm is usually dependent on whether we actually have the data. So timelines are based off of how quickly can we get the data from the client to start working on at least an initial hypothesis? A lot of time gets spent also on just the cleaning and really distilling and building sort of a single source of truth which one can go off of and the team can go off of.

So you mentioned data cleaning there. And I imagine that if the client knew how to get the most out of the data themselves, then they'd probably send you the clean data set. But that's the reason they're hiring you, right, Is because they need help working with their data. So I can imagine that you do have some fairly gnarly data sets to deal with. Just related to challenges with data, how do data privacy and security issues affect you? So if your clients have to share data with you, do you run into any roadblocks around privacy and security?

Well, that, that is always the concern and that is always something that we keep an eye out for in terms of how we work and the kind of system so that we're not violating any, any policy, any, any regulations. I think GDPR is, is one huge regulation that we work with. So like, let's say working in Europe is working in or anywhere. Actually, as a matter of fact, it's paramount that we don't work with BII data and, and with that also understand that the analytics that we build is ethical. So there is an ethical component to it. It's being responsible and it's to ensure that we were not introducing any, any bias in any systems that are getting built. So we'll work with legal teams, we'll work with HR teams, both of the client, ourselves, our, our own teams to ensure that we're, we're working and in compliance.

Okay, that's really interesting that you make an effort towards having these ethical systems and do your data analysts and your machine learning scientists, do they have any training or expertise in how to make ethical models?

We've definitely been going through that over the last few years. That change has come through in terms of both as we recognize how the data that we already start with might already be affected. How do we actually help our clients in dealing both with what's already been recorded versus also ensuring that there isn't a bias that comes through the system. And in order for us to do that, we, we work with our data analysts, our data scientists in vetting those models and we're involved in the quality aspect of it on an ongoing basis.

And I'm curious as to what your customers and clients want out of the projects. So my experience of consultancy is that most of the time they just want to know how you're going to make them more money. But occasionally you sometimes get someone who's technical and they really want to know all the details of how your methodology works and things like that. So I'm curious as to your experience of what your customers actually care about as the results of the project.

I've encountered this quite a few times where clients want to really deep dive into the modeling aspect of how does this work and let me bring the expert from my team. And frequently that's an amazing experience because you can quickly nerd out on models as well as model specifics that helps us in socializing what's been developed. So it's not just that we've developed something only for the sake of the client gaining an upside, but it's also there's a technical component in helping them establish the longevity of such said solutions with their teams because ultimately the client teams will own such solutions if they've not been co developed with the client. And having said that, I think the more curious, I think we have folks from clients, I think the better it is because it also helps us understand that there is a team that can catch the solution itself or can be a part of that solution building Process.

Okay, so it seems like that involves some sort of level of skill transfer to the client and I'm curious as to how that happens in general. So for example, after your project is finished, how do you make sure that you don't just sort of leave the client to deal with whatever you've given them on their own? Is there some sort of maintenance that goes on after a project's been completed?

We work on helping the client either stand up the capability. So like personally I've helped clients stand up an entire digital organization, help them stand up their data science organizations. So it's both capability built, but then also the training of teams. So we'll do let's say two in a box, a solution where we co design the solution with the client and along the way we clean up the client's team members. As an example, on the digital twin front, like I've personally done that, where we've trained client team members while developing the solution so that the client actually had an entire team to continue working both with the solution that was developed, but then develop new solutions, new digital twin solutions. And so that was capability building, both from standing up the team, but also training up the team. That's frequently the way that we'll go about. But those are definitely larger projects, larger programs on the shorter term, on the sort of smaller scale projects, we document everything, we hand over materials to clients, we organize training sessions so that can ensure sustainability for the client. And of course we're around for our clients to be able to answer questions while they go off on the next chapter of their journey.

And just circling back to something you said earlier about you often have quite strong time pressures in terms of delivering projects. So it seems like you must have a few tricks up your sleeve in order to be able to run projects efficiently. So do you have anything you can share with us?

So I'd say like on that front, I think we know for a specific industry what to be looking out for, what are usually the areas where clients might be suffering. Then of course our own relationships with the client. We were able to get to it much quicker. I think on the project management aspect of it, it does become that we know what the typical risks are and where there might be slip up. And that's already a part of the training curriculum. So anyone who's becoming a new manager will go through the basis of getting that kind of training. As you're an associate and going on to becoming a manager, you're receiving those trainings along the way as well. And it's Important, you, as an associate or as an analyst in the firm, you're managing your own work, but as a manager, you're managing not only your work, but you're managing your associate analyst work. And quickly those trainings start becoming very useful because you're taught in how to structure your work and how to structure your project.

Okay, so having that sort of structure and the template, it sounds like it helps you reuse work from one project to the next. That's something I'd like to get into in a bit more detail. So how do you go about ensuring that you're not having to reinvent things from scratch every time you start a new project?

So in order for us to do that, we'll capture the stories, right? We capture our stories of what we've done in previous projects and are they similar in nature to what we might be working on currently or we might be working on in the future. We use that to build our base of how the project went. And it's. It's building off of those stories to be able to go ahead and say, this is our base of understanding on how that went. Usually these are common pitfalls for this kind of an engagement. So, like as in, if we take predictive maintenance as an example, we know what systems one will be working with. What are common pitfalls of predictive maintenance in the forestry industry or in the airline sector, or in the automotive sector, and how do you help a client through that? So building our base of understanding, going off of what we've already seen before is usually helpful. Speaking with those teams. So opening up, like, having a much more collaborative channel on what works and what doesn't work.

And if you're running multiple projects at once, how do you juggle the context switching from one project to another?

That's always good. There's always a challenge in that. But there's also the thrill in being able to juggle a few projects. Why I say that there's a thrill is the fact that you can see the movement on all these different projects at the same time, which is interesting. I think the context switching can be a problem, especially if you're new and if you've gained the experience, I think you've got some years under your belt, you start becoming better at it. But there's still always going to be a slight challenge around context switching. And we try to limit, like on how many things that one can pick up as an individual because there's only so much time in the day, there's only so much time in the week. That you can devote to topics. So how many topics one can pick up and make a meaningful progress, you decide based on like where you're at. And as an associate coming into the firm or as an analyst, you're focused 100% on one project. As a manager, you're focused on 100% on a project. And thereafter it starts becoming a question of like how much time one needs to put where and back.

Near the start of this, you mentioned your project in South America about working with forestry and having these sort of AI solutions which sounded quite high tech. So maybe can you talk a bit about what the technology and that involved?

So this one was more around helping the client understand and build a uniform product. And the uniform product being pulp production. So this is pulp production from their own sustainable farms. How do you understand the biochemistry of the wood that comes in to be then able to say what's the product quality going to be? If you could predict that, you can then preemptively adjust levers to essentially ensure that you're getting consistent quality, or you can ensure what mix of wood goes into the creation of pulp to be able to ensure the uniformity of quality. Why that was interesting was because we had to study what it meant in terms of the internal cell structure of trees and what sort of is the moisture content to be able to then say how much of it is going to translate into a bad quality product versus a good quality product. So it's fascinating because one wouldn't ever think, or we didn't really think that that's the key aspect. But again, working with the client and the client's data science team, we were able to gain that understanding because the data science team there had already been working on something like this to try and figure out what's the internal cell structure and moisture content. And we were able to leverage that to take it further on the product quality side.

So that's fascinating. I realized that question was a bit of a jump from talking about project management, but what that sounds like is you have a very specific solution to that problem. I'm curious as to what happens when you have to build technology for one specific project, because it's quite a niche specific thing. But then you have all these other projects. I'm wondering how do you juggle with the different technologies in different projects?

I think that's part of why one is. I think as a data scientist, as a data engineer, as an engineer, I think in working in consulting, that's the part of being curious is we need to juggle with frankly, many different technologies on an ongoing basis. I think the one key thing always is, okay, high client, what are the different technologies do you use today? Is there room for us to push either the envelope in terms of bringing new technologies to the client, but we also need to take stock of the situation on where the client's located. Do they have access to these skill sets in their talent pool, the talent pool that they're likely going to hire from? And so we have to take stock of that situation before we go ahead and arrive at technology specifically. But outside of that, I think it's codifying what we can at a higher and an abstracted level from each of these solutions that get built. So as an example, the forestry example I was sharing, we were able to take it to an abstraction level where we could codify saying, this is what happens when you are trying to solve a product quality improvement problem in so and so specific industry. And that's like continuous production. Now I could take that and apply the same concepts in a different industry, which is still continuous production and now just use that as a starting point of the understanding, but explore the technology sets separately, which are going to be more relevant for this new client.

And you've mentioned curiosity a few times throughout this. It does seem like continuously learning new skills is an important part of being a data consultant. So can you perhaps tell me a bit more about how you go about or how your team goes about learning these new skills?

I think both. The research, Richie, is like, if you encounter it, researching what might actually help get us off the ground quickly enough. So as an example, let's say you working with a data set that was sent over by a client to us. Now this came back with a consultant on a secured drive and we weren't able to travel. This was a different geography. This was a sensor based data. So it wasn't pii. But looking through that data, like how do you handle that? That was. We didn't have any context on what those files look like. We had absolutely no connection with the client. Given the remoteness of where the client was located, we didn't have access to them. And so how do we look through and sift through the data and even understand what those files contained was an interesting challenge. But one of the consultants on the team was able to quickly open up the file in a UNIX environment, peek through on a couple of lines and then be able to decipher that. Well, this was, this looks like a certain data type and then research it and be able to download essentially the software to be able to go ahead and start reading those files. And it was that or, or being able to. I think in another example, it was, okay, this file actually has delimited values which the delimiter is unknown to us. And quickly being able to write a Python script that would do on a line by line basis, output the data cleaned up with a delimiter that the whole team could work with.

Okay, so this sounds like a lot of detective work and problem solving in cases where you're not really sure, you don't have much context and you don't really know exactly what you're doing. But just having that enthusiasm for going in and having a look and seeing what you can get out of it, that seems like a really important skill.

The more tools you have on your fingertips, I think the better that journey looks like my own personal journey. Having the familiarity with UNIX based systems or the UNIX toolkit, I think was fantastic. Just being able to open up quickly something. Being able to code up using regexes and quickly clean up files, I think that was an amazing skill to have learned at one of my previous employers. How regex works. Right. And that seemed to me like almost like a Swiss army knife for a data analyst.

Absolutely. Regular expressions, regex is. They're perhaps one of the most underrated skills. It's like it comes in handy in so many different projects. Absolutely. Well worth learning. I agree. Do you have any final advice for people who are interested in data consultancy?

I'd say the future is really bright and there is high demand for anyone who's looking to join the data journey itself, especially on the consulting front. I think we're just getting started. I think the whole world is getting started on the consulting side, so just head on over. We're always looking for amazing people.

All right, wonderful. That's a great note to end on. Thank you very much for your time. It's been really fascinating stuff. I'm getting excited about consultancy now. Thank you.

Thank you, Richie. Thank you for having me. And thank you everyone for tuning in.

You've been listening to Data Framed, a podcast by Data Camp. Keep connected with us by subscribing to the show and your favorite podcast player. Please give us a rating, leave a comment and share episodes you love that helps us keep delivering insights into all things data. Thanks for listening. Until next time.