okay? Sure. Um, early on thistles, my data science to my second career. My first career was in animation. Eso coming out of high school. I was always very interested in a couple of different things. One was programming, and the other was art Andi. I kind of picked more of the art route. I went to art school, got into animation. Um, and as I started my animation career, I kind of came back to programming on started, uh, working towards a concentration in an animation. That is, ah, kind of a mix of art and technical aspects. So a lot of programming, Um, and after a fairly long career animation, I was probably in the worked as professional for about 13 years. Um, kind of started to get bored of the day to day work, was doing a lot of the same stuff, solving a lot of the same problems. Uh, kind of started looking for something else. So this is maybe, like, 2018. Um, I'm just kind of fun. Data science through ah, Google search. It seemed like a good combination of the things that I still wanted to do from my my animation career. so namely programming problem solving, Uh, and had some new things that I was getting more interested into, Like econometrics. Eso It seems like kind of a good combination of those interests. Um, I started doing some online classes, some free online classes just to see if I was interested. Uh, turns out I was so from there, I started looking at the boot camps, which is how I found the Medicis data science boot camp, which is where I went through. Um, so I did that boot camp in 2019. After finishing the boot camp, I did a little bit of TA work for the same boot camp, and then I got my first data science job, which is where I am now.

eso. I'm kind of, ah, dual role. They're pretty much everybody that works for P. A. Consulting is a consultant on I'm no different. I just happened to be a consultant and a data science specialist, so I'm usually on jobs that involve some kind of data component. Andi, I'll either be working a Southie soul data science specialist or on a team of data science specialists. And in that aspect, the data science work that we do isn't particularly different from the data science work. You probably see, um, at other companies, I think the main difference is that we also have the consultant aspect. So, um, it is our job to kind of come in, evaluate where a company is in terms of their data maturity, Um, and in really what? Their needs are on kind of the current status of whatever data they have that they want us to do something with s. So we have to kind of go in, evaluate all that interface with the client, then usually build out some kind of project. It can either be the output from that can either be, um, some kind of dashboard or tool that we live with the client. Or sometimes it's a simple as well Take a look at your data and we'll make some sort of recommendations or answer some sort of questions that you have based on that. So just to summarize its its normal data science work. So a decent amount of data engineering, Theun data visualization, some data modeling, um, our predictive modeling. And, uh, a good portion of that, however, is also speaking with the clients, making sure that they understand, Um, what we're gonna do. The limitations, I think, is a big thing. We have to communicate of the techniques. We're going to use the limitations that air presented by their data on, then kind of explaining the results to clients that are obviously very well versed in, um in their specific sector, but may not have a lot of experience with data science techniques and various data science metrics. So it's kind of like translating those those things that are important to us as data scientists when we look at the data but translating that into, um, language, that is easy to understand. For non data science, peopleYeah, good question. So I'm gonna kind of combine the priorities and pain points because I think they kind of go hand in hand, especially with, uh, with consulting work. So one of the biggest things that we deal with is communication with clients, and a large portion of that is managing expectations. So we often when we go in, we don't know exactly what state the data is in that our client has Often the goals are a little bit unclear. And these two things are drastically like changes in what we find there could drastically affect the timeline of a project. So when we go in, we kind of have a rough time scope toe work with because it's based off of what? We're what PPA is billing the company, and we kind of have to fit in, uh, within that framework, we kind of have to fit in how much we can actually get done. And if it you know, there's some wiggle room there. So if if we come back and say it's we're only going to be able to dio, you know, give deliver x amount In the given time, we might be able to extend the time and charge Mawr, or we might not have wiggle room in the budget and we might just have to scale down what we're trying to do. So communication about limitations and expectations is definitely one of the top things there. Um, another priority is, uh is making sure that we actually communicate properly, what it is that we're doing and what are important metrics for success. So I kind of mentioned this in the in the first part of the question. A lot of our clients are pretty new to analytics or, um, are not data science people themselves. So they they have every business has their own set of KP eyes that are that are important. Andrea's data scientists obviously have our own metrics. When we're dealing with things like models looking at data, looking for looking for different issues with the data, we kind of have to bridge that gap. So we want to. Ultimately, we want to make sure that we phrase, uh when we're talking about the the data science, we want to phrase that in terms that relate it back to the KP ice that are important to our clients. So we wanna make sure that we're not just talking about Oh, this model has, like, a has a great r squared or or the the accuracy and precision O R sorry, precision recall or look really good with this other model. That's not gonna mean anything. We need to equate that Thio. You know, this model is, uh, you know, well, has the potential to generate this amount of money in savings. Uh, we need to state the business case, uh, in order to properly communicate with our clients. So I would say that's the second priority there. Just to sum up is always relating what we're doing back to the business case on the third priority is really a time management. Um, so this is this is less on the dealing with the client side. And this is Maura as a consultant. Um, a lot of consulting is kind of coming in, and to some degree over promising what you could deliver on delivering it anyway. So, um ah, lot of times on our jobs, we just have to be super efficient. Otherwise, even being super efficient, there's gonna be some late nights, long hours while while we're finishing up client work. But, um, inefficiency is just not an option we have. We have to be, uh, basically like we have to be really focused on what we're gonna get done and how we're going to get it done in the a lot of time. Because unlike other in other scenarios, before I was a consultant, I worked obviously the non consulting worlds on there. You know, sometimes they're hard deadlines that you can't pass. Sometimes the reality is you're just not gonna get something done in the allotted amount of time and you could get an extension. You kind of re prioritize or rework the timeline in consulting most of time. That's not an option. So you just have to figure out how you're going to get something done in the a lot of time and then execute So, uh, to sum it up, the top three priorities, let's say, are communicating expectations and limitations with the client, making sure that when we communicate with the client, our progress we're communicate were communicating in the native language, the native metrics and KP ice that matter to them. And the third part is having good time management so that we could get everything done for client project

okay? Yeah, This could be pretty much across the board. I would say if we if we have the option, our data science stack is pretty similar. Thio what you see anywhere else? Um uh, we we don't We're not tied to either python Or are some of my coworkers use our I've been programming a python for a long time, so I use python specifically through the anaconda distribution. Um, I use a lot of the common libraries, like, uh, see, born in that plot lib for for visualization, obviously using pandas and SK learn, Um, in terms of B I software, Uh, that can be, ah, little more dependent on what the client actually prefers or has licenses for. So, uh, I generally end up using tableau just because a lot of our our clients also just they just happen to have and use tableau. But I've also had to branch out and learn a bit of power bi I and looker. Um, it all really depends on what the client's preferences. I find a lot of times, the clients that we work with, um, they are set up with sequel databases. So Microsoft Sequel server is a Big One. So that's usually where we get the bulk of our data from. And, like I said on the visualization and tableau is probably the most popular, Um, in the middle there, that is usually kind of a soft spot where our clients are usually not as, ah.the kind of like soft area in the pipeline is usually around predictive modeling. So our clients are usually they usually are, not they. They're usually at the data maturity stage where they they have the ability to store a lot of data. So they're using a sequel, usually a sequel database, or some some other kind of of data framework. Um, and they usually have some kind of data visualization capability. Uh, but they don't really have any, um, any kind of pure programming capability in the in the middle area there. So not much in terms of predictive modeling. Um, not much in terms of automated processes. So that's kind of where we come in. And that's kind of the gray area Where, uh, I personally, we'll use Python, um, whenever given the chance. Thio whenever it's an option. But it really doesn't matter what we plug in that right there. It's our python. Whatever you're most comfortable with, Um, the big key here is, uh, if the client wants to be able to pick up that work later and build upon it, um, we might have to work that out with them so that if they say like if they think, for example, Oh, we have somebody on staff that has learned are in the past. Uh, they might be able to pick it back up again. Then we're probably going to be writing everything in our so that we can pass that work off. Um, otherwise, python tends to be more popular just because more people use it. It's more common to find python developers, so that is the only sticking point there is. If the client wants to eventually pick up that work after we leave, they may have a say in what we what language we used on. That's obviously going to affect everything down down the line, right? So if if we're doing some visualizations, uh, if we're if we're using our, that's gonna mean probably using g plot versus If we're using python, we're probably going to be used using, like, combination of Seaborne Matt plot lib. Um but otherwise, uh, it's pretty much our choice. And like I said, the beginning part. So the data storage part and the usually the data viz part is kind of up to the client. It's depending on whatever licenses they have, or whatever the software preferences are