office-hours-nov20-2020.mp3

[00:00:06] Oh, yeah. What is up, everybody? Welcome. Welcome to the @TheArtistsOfDataScience Open Office Hours. I am super, super excited to have all you guys here. Holy shit, man. The waiting room is popping off. There's so many people joining in today. I'm so, so excited that you all are here. And I cannot believe that we have to Data robot himself. Ben Taylor, what is up and how are you doing? Hey. Hey, man, I'm so happy you can make it. Oh, man. We got Tom AIs here. And Tom, we've got a while. Monica AIs. Oh my God. And Giovana. Oh, my God. This is this is awesome. And this is super, super cool.

[00:00:46] I can't believe how I wanted to say it's been so long. I was having been withdrawal. So I'm so glad you're here though.

[00:00:54] God, it's been two days. Three days. Yeah. We were on a panel in Africa. It was fun. I saw that man. How did that go? Good.

[00:01:03] We we were enamored with ourselves. Let's just put it that way. Yeah.

[00:01:08] That's how it it seemed like a great discussion. Seemed like Alexa and Ben and I were pretty well aligned. So that's always fun. No fisticuffs. But, you know, Ben, I was I was laying in bed this morning. I was for some reason, I was replaying some of our comments and I realized, you know, that fear you have that you made this highly waxed philosophical comment and you really didn't connect it. Back to the point of the show. I was thinking that anyway, this might be fun to say sorry. I don't mean to take over the show, but no, that's good.

[00:01:42] I mean, you woke up in bed thinking about it. That's that is cool, man. Yeah.

[00:01:47] It's just that our affection, it takes so much focus to become a thorough full stack Data science. I'm not claiming to be one. I'd love to become one, but the amount of focus it takes is like being in grad school. And it's very rare in your lifetime that you get to afford that much focused on anything after school.

[00:02:09] No, absolutely. So I guess how would you define a full stagnated scientist? I know what a full stack person looks like. I seen that meme floating around LinkedIn today feels that person cuts here. Kirks develops vaccines and a number of other things. But what is a force that data scientist?

[00:02:26] I will I will answer first and been promises to fill in my holes. So imagine that you are a web developer, RSS developer, let's say. And so you're doing the back in the front end and you're going to add some Data science tools in the mix. Maybe you've got some launch pages you're experimenting with. And they said we don't want to go get a Data science without that. You do the one armed bandit type stuff for figuring out which of these landing pages is most effective. And then you go on ahead and do all the great data visualizations and all that. Well, I think where we all landed was to say, if you're a startup and you have as one of your primary members in the startup, a pulse, that data scientist that can prove out the Emdin solution, that's practical. But if you really expect to be successful with just one person on a large scale s in the long run, that's kind of insane. And I'll let them fill out what I didn't.

[00:03:25] So, yeah, I agree with a lot of that. I think the fun thing my experience has gone from helping get a Data science shop, getting a Data science job, being a Data science manager and then writing payroll checks to Data scientists. And that kind of takes you to the dark side where you want to fire everyone all the time. So I feel like there's definitely a sense of urgency. So I think the biggest fear is just urgency. So just get things done. If you have to go search some stack overflow for some flash scrapper or some little missing part, do it the front end engineer. So make fun of what you did. But the fact that you did it in a day or two is super helpful. So like, can you build an app in a sprint even if it's embarrassing? Even you have to bust out some HP like I started as a PDP developer, so all the printed developers make fun of me. They're like, oh, gross, you still use that crap. But that's what I started with. So I think a full stack. I think of urgency.

[00:04:18] Yeah, I just do it delivering, getting stuff done. And you don't even need to necessarily be, as you know, the data scientist at a startup, you can be the data scientist, the first data scientist in an organization starting up a data science practice from scratch and having to to have all those skills.

[00:04:34] And that was the experience I had at HireVue for the first year. It's just me. And it was really helpful to have school experience because you didn't have to. Mother, may I through the engineers and you could just kind of go pecking around and get what you needed were if you had to work through the engineering team, everything would have slowed way, way down.

[00:04:51] Yeah, that's so awesome. And so we're here to help all these wonderful aspiring Data scientists up and coming Data scientists. I just want to shout out a few people that are consistently in office hours. That's Kristen. See, Ashin, thank you guys for showing up. We got a ship here from last week and Karen, of course, is in the house. Thank you guys for showing up. Again, a very special guest today. We got Ben Taylor. We got Tom. I've got Dave Languor. We got Monica Royal, Giovana, Sasha, whom I think there's so many people on my screen right now. This is insane. Last week was stacked. Wow. Yeah. Yeah. Then when you guys listen to this replay on the podcast, go to the YouTube channel and just see, like the giant smile on my face when it's so cool to see everybody here. So, yeah, let's let's flip it to the to the students here. I mean, they use that term liberally. But if you guys got any questions at all about breaking into Data science, about project ideas, career advice, whatever, you can see here that there are a ton of people to give you some great advice and share that wisdom. So whoever wants to go first, by all means go for it.

[00:06:01] Share mistakes to know what not to do.

[00:06:04] What's your topic number one not to do besides deleting production tables?

[00:06:08] Oh, man. Actually, I'm more sensitive to mistakes. I keep making time management just sucks.

[00:06:17] Just like time management. The constant battle, especially when you're you have meetings. I'm not used to having so many meetings like now in my new job, I get triple booked like I've never been triple booked before. Like you like it's constantly like, OK, not going to this meeting. And so yeah. So I'm very thing that's up in mine is how do I sharpen the saw every week. And that becomes very difficult by sharpening side.

[00:06:37] You mean keeping up on text ax or any skills.

[00:06:41] Yeah. Keep it up on Newtek Stacks. I do want to throw myself under the bus so throw myself under the bus. One of the mistakes that I have repeated too many times my career is if the results are too good to be true, they are always wrong. And I've had like this has come up like so many times my career. You get a modeling like, yes, the CEO is going to love this or this is going to change everything and the customer going to love this. And you have to be for people that are more seasoned, they're never optimistic. Like it doesn't matter at the model results are they're never going to smile. They're never optimistic because they've gone through the pain of skewed results are like all the reasons why you you were misled with good news.

[00:07:19] Yeah, something like that happened to me earlier this year. I had developed a model that was just performing amazing. Like I was legit clairvoyant with my response to it, with my results, rather. It was just so amazing. And I had to approach it with a bit of skepticism, like, actually, wait, I can't tell the future. So there's no way that my model is good. And it turned out that I had accidentally employed the wrong cross-validation technique. I should have been using group called cross-validation, but I was just using regular Caple cross-validation. And after making that small little tweak, my results were kind of more in line with what I was anticipating when I was expecting. But I think it's I think that's the importance of just not just always learning and always reading up on stuff, but just doing things right, like doing projects and getting exposed to how things are happening in in the real world with real Data. So you can be skeptical of results like that. We got Carlos here as well. This is this awesome.

[00:08:20] Carlos literally typed holy shit in the chat kubernetes.

[00:08:24] Yeah. So, yeah, dude, this is awesome. So I know people might be shy to take the floor, so I will start actually Harp.

[00:08:34] Yeah. If you, if you don't want to have to do the general start, I kind of have some things I can jump in with. Absolutely.

[00:08:41] I won't take too long but I haven't. This is probably on the week too if we plug it in here and really probably only last month or two plugging into the community in general and just being amazed by how much everybody's willing to share.

[00:08:54] And this is a perfect example, looking at all the people here and saying, oh, just a quick background for my questions, just kind of give in perspective. It's you know, I'm a business guy who's gotten into Data within probably the last one or two years. And so, you know, when people talk about communicating with the business stakeholders, that's me. I'm that guy at my company. We do the commercial management, things like that in my in my group. And I'm also just really into Data. And so probably more recently, I decided I need to get more serious about this. I look around a company like my own, which is a fairly large energy company, and I see that we have done a pretty good job in the last one or two years of collecting data from our employees of our assets and things like that, talking about Data, Data, things like that. We have a good repository of that kind of information now. And so we don't really have a lot of people in-house, let's say that can do stuff with that just yet. And we do have a few people that are interested in it, including myself, and we probably don't have the tools yet. So one of the things we're trying to do is scale up. But if I think about trying to spin up like a minimum viable product or trying to improve things out for senior management at a company like mine, people who have done that before and several of you here, I'm sure, have how? I think about approaching a minimum viable product at a company where we have good Data, we don't necessarily have the Data culture maybe at the highest levels of this company. We haven't understood yet how to leverage that.

[00:10:16] Stay away from machine learning, Harp.

[00:10:19] If I can, I'd like to just. Yeah, Rosemount and I think Dave and I were having some good comment support back and forth from his post on this. But I like to liken it to a control system design. We have a feedback control. There's so much wisdom in that for how we interact with our business counterparts. Christian and if you study that those systems, you realize if you don't get frequent enough feedback on your current state, you compare that to the state you want to get to, you can get into large instabilities. And I think that is so true. I've tried to steer the comment space in LinkedIn away from you have to have good business acumen. Well, you might even have better, better business acumen than the business people you're working with. Probably not. I'm just saying, even if you did, what if they don't think you did? What you're trying to do is build that relationship, find out where they're hemorrhaging, find out where Data can help them and give them just really small, tiny releases, frequent illustrations of where you're going and why? Because they'll let you know right away if they understand what you're doing, if they appreciate what you're doing, if they think it's really going to have a chance of helping them. I think the first key for us to give Data vandalism into the next stage is to win these people over and understand we're not earthlings, all of us on the screen. We are barely earthlings and we don't we don't speak. Our native language is like our business counterparts do. So by just interacting with them more and showing we care. By the way, I'm being long winded right now. Just replace everything I said with a four letter word care by sitting in front of your business counterparts. And if you think this is going in the right direction, I think it is, but I care more about it. Do you think this is meeting your needs?

[00:12:17] And I'd love to hear what others think about all that, but I just think quit trying to develop it from, you know, to the final version, show several milestones along the way, even more than you would be comfortable with upon that care thread.

[00:12:31] Real quick. So one of the problems with machine learning, there's too many things to do. And you probably look in that question when you look at like what's under the hood, you're like, oh, wow, look, I got this Data that did at this Data. And a lot of the vast majority of our projects fail, but they don't fail because of the technology. They fail because you're working on the wrong problem. And in the spirit of caring, someone's paying payroll, they're not this isn't like a postdoc position where you can just do Blue Sky Research all year, no value. And so my recommendation and I'm speaking from a history of mistakes, I've made a lot of mistakes here. Try to shoot for like a one month target. Can you find a winner in a month with some analysis? There's a really good chance you actually can't put budget that time in your normal working hours. You might work a little bit late, come in early, maybe work on the weekend. And then when you have the meeting, the temptation is to show all your work. I tried this, I tried that didn't work. And I could teach you a little bit of statistics. Don't show any of that. Just get right to the KPI, don't teach them anything and just say I found a when it's significant. And this is how we can validate it with experts, not with data science, with domain experts, technicians or whoever actually knows the data. This is how you can justify what's happening and what you'll find as you get one of those wins. And you're now the superstar. You get one of those wins, money's coming your way. They're willing to take bigger bets. Maybe they'll let you do a six month project. And I've seen Data science teams blossom. They have to have the first one on the idea of doing like a six to 12 month moon shot. People hate that from a business side. They hate it.

[00:13:54] And I think then correct me if I'm wrong, what Ben is trying to say is what you can deliver to make yourself a hero is so much lower than what you're probably thinking.

[00:14:04] Yeah, if I could I could riff on this a little bit as well. And I'm speaking from Hardwell and experience, which means I screwed things up the hard way and learned find just what Ben said exactly. Under promise over deliver. And the easiest way to do that is find a manager, find out they're compensated and find some sort of data analysis scenario that helps them make more money. Bonus promotion, whatever. That's what a win is. That's what a quick win is. And once you do that, you're off and running. People love Data at that point. They love it.

[00:14:40] Ok, ours are the new CPI's. So like Dave said, find out what they're OK is, find out what they're actually getting bonus on that quarter and make that no bigger objectives and key results for those guys won't do it.

[00:14:50] Ok, ours are definitely. Get the book by John Dugher. Measure What Matters. Excellent book. Creation is the answer. Your question.

[00:14:58] Yeah, and then some. I mean, you know, probably some business oriented guys. So a lot of that makes sense to me. I've been on the.

[00:15:03] Receiving end of being told about all the inner workings of a thing that I really don't care about, we in our company especially have a pretty good feel for for the big things that can that can drive our profits one way or another, especially an asset based company. A lot of it's based on cost and expenses of driving those down, especially in this environment. So that reaffirms some of the things that I thought and also gave me a lot more to bargain for, to think about.

[00:15:26] We should get some. So I'm going to go clockwise based on what's on my screen.

[00:15:32] So next time I see any questions or the ones that don't, you know, no questions, I just coming out really for sure. Amanda is always welcome as well. Sasha, how about you?

[00:15:43] Yeah, I have a question for those who are just entering the data science and analytics field for research bootcamps graduates. Let's say there is a common notion that data analytics is a reasonable route of entering data science field. So I was wondering if what do you think about that for about people starting as data analysts and then transitioning into a data scientist, or is there a better way of doing that?

[00:16:08] So leaving the boot camp and starting as a data analyst before getting the data science title, this is interesting. This is a side note. So I have had beef with colleges that graduate their data scientists and a lot of them end up as analysts. And I see that as a failure of like a formal college because they haven't prepared them enough for them to go land one hundred and ten, twenty, thirty forty thousand dollars job out of school. So they have to go. Landed an analyst position or is it doesn't really answer your question. My hope is that the boot camp would be good enough. They get a Data science title right out of the gate and it's competitive pay and competitive pay is like one twenty starting and up. Unless you live in the Bay Area, that's one fifty or I don't know what the numbers are these days.

[00:16:49] Yeah, I do live in a Bay Area, but it just seems the industry seems so saturated here and I've been searching for a while. So I was wondering, would it make sense to just adjust my expectations for a while, especially with the job market during the covid being so wild?

[00:17:04] I think this will change your title too. Sorry. So if you land as a data analyst with the right company, they will change your title the first year. If you're delivering the value, that might be something you can negotiate. Sorry, I cut someone off.

[00:17:15] Oh, I was just going to speak because I actually was in this position last year. I did a boot camp, I did the Data incubator and I actually didn't complete the program because I was able to get a Data science job. And I think the challenge was that the mini boot camps, they they position themselves as like you do a boot camp for seven weeks and you'll get a job afterwards. And I found that that's just really not the case. It's more so they give you the skills to learn, which I definitely need at that time. But they did prepare for war like the job market or like the interview process. I think the interview process for like a Data and compared to a Data, scientists are completely different. And so for me, I was already working part time as a data analyst for a separate, separate project. And so I was able to leverage that into going to a Data science role right afterwards. But I think the main thing is like for me personally, I think it doesn't really like the title. Like, as you alluded to, like salary wise, the title does make a huge difference, but especially for your first job. I want to focus on the title. I will focus on what value your role would drive to leverage that into the Data science role. And so for me, like I got my first job as a data scientist and that was just timing and luck. I met the person at a conference and I convinced them. But if it was a data analyst who also took the same job because of the type of work it was doing, it well positioned me well to go into a data science will want to rope in either Monaco or Giovanni.

[00:18:57] If you guys have any input into this question, I'd definitely love to hear from you guys.

[00:19:01] Yeah, definitely. To add on what Mark said, I actually have something against titles themselves. In some cases they're not what they seem.

[00:19:11] So you would apply for a Data analyst role, but then you are required to perform things that a senior data scientist would be performing. So I would actually focus on the requirements and the duties of the specific roles versus the titles themselves and not and also not trust those role descriptions to the TI. I've been caught up in that as well, whereas AZO you'll be using such and such tools or such and such qualifications and none of that actually existed. So during the interview process, it would be very beneficial to ask them up front, like what actually will be my day to day activities and I'll jump in.

[00:20:01] A lot of people don't realize that like. You're going to switch jobs like you're not going to stay wherever you're at. So get one that gives you really good stories and then just make the switch and don't tell them what your salary was. You can double your salary if you just don't leak stupid info, like make lots of grand narratives, crush it and then omit whatever you want to admit. So that's I want to give you the shot, like especially like I hate to say it, but, you know, if you're not the right race, you got a funny sounding name like you're going to get you're going to get screwed over. So, I mean, you go do what you want to do. For me, it's like you is money, good stories, big clutch stuff. I'm actually looking right now for a post that thought earlier about like an online pseudo master's and Data science like that. And all the bootcamps I've posted in the chat.

[00:20:45] I really like what you said, Carlos. I think you have to fight for the job you want three years from now. And taking that job today, don't think about the job today. Think about the job you want three years from now. And the job you have today might be a fantastic catalyst to get you where you want to be.

[00:20:59] I was hoping. Oh, let me. Whoever is going to be OK.

[00:21:04] Thanks, Tom. So this might be a bit of a heretical view coming from me. Not a shocker, but here's the way I think about it. I think Monica is exactly right. You have to be skeptical of job descriptions. You have to be skeptical of titles. But here's the way I look at it, and I'm coming from the fact that I get a lot of gray in this beard. So I've been around the block, OK, as much as Ben. But I'm older than Ben, by the way. But here's the notion. If the data scientist position requires you to regularly write and maintain production quality code, you're probably not qualified to do the job. And the reason for that is if you graduated with a bachelors degree in computer science from Caltech or some other high end school, you are also not qualified to write production code of any kind. That's speaking from 20 years of software engineering experience. So also keep that in the back of your mind as well. Can I really do the job? Because the worst thing possible is to learn that data scientist title, get that big fat paycheck, get on the get on the floor, work in the trenches and you can't actually do the jobs at the level of expectation that's that's just going to shoot yourself in the foot. So I would always think about that. And Monica's point, if you're an interview, ask them, are you expecting me to write and maintain production quality code as a primary aspect of this job? And if the answer is yes, consider what that means.

[00:22:22] I really think this discussion is moving in an excellent direction, no pun intended. There they but truly an excellent direction. You're suppose that was your cue for a guitar riff? Just for. But anyway, I like to summarize it this way. It's a care level for this discussion. It's being definitely do have an Monaca. I really like the way you were posting. Things always seem to be a Data science or at least becoming more and more data scientist. And then, well, if you are a data scientist, you do Data science work, do the role you're in well and enhance it with Data science before long just to keep you, you could negotiate. You need to call me a senior data scientist because that's what I've been doing, been doing, or I'm just going to have to take the software I have in hand. You decide really if you concentrate on being a data scientist and literally loving Data science, I think what then? And then David Sanger and Carlos, excellent words. In my opinion, it's going to happen.

[00:23:31] So I love to open it up to either Jeevana or Kate, keep stretches in the house, these Delmon. So we're talking about the interview process there. So hopefully that answers your question. If not, let me know. And by the way, if anybody has a question, just please type it right into the chat. That way we can keep you kind of next in line. I'd really appreciate that kind of keep things going. But I mean, there's a good portion of the Data science interview process, obviously, that is all about data science, but there's parts of it that are all about soft skills. So can we get your insight, Giovana, on the need for soft skills in Data science kit as well, if you guys can chime in.

[00:24:11] Thank you, Harp. Yeah, I think how you present your Sachar, how you present your your skills is an important way of how they are going to consider you for their position.

[00:24:29] I think this is the most important part of your presentation. And and I think that you have to forget when you are doing an interview is that you are a beginner because you do have to avoid to say, I'm a beginner. I don't have experience that you have to eliminate these words, these phrases, because you have experience. You you have do a lot of projects, I think, because in a boot camp you have. You have had the opportunity of work with other people and you have maybe you have been part of a team, you have to tool to do a lot of project and you have that from your projects, because when I did my Google, I had a lot of mistakes. I get stuck a lot of time and I and I feel this frustration and I need to ask for information. And when you need to ask for information, you need to look for information. You are training. That's part of your experience. So that the data scientist, because when you are going to work as a scientist, you are going to connect with other people. You are going to come in. You need to communicate with other things. You need to ask for help. You need to ask for advice. So you are training yourself his skills to be prepared to be an excellent data scientist. So I think you when they said, what is your experience? You need to do well and talk about your problems, the faces that you have to do when you are building your first model, or maybe the things that you learn when you do your first deep learning model.

[00:26:26] You can talk about that because that is your your experience. That is the things that you have to present. And then and one thing that I always said is if you have projects, you are better than a lot of people that think about and they have a lot of ideas, but they don't finish any. So a person who has projects finished that they have the repository that can show things that they have done. They are valuable because they are able to finish a project. So you have to present yourself in that way that you are a professional, that you have experience, maybe you don't have experience in a company, but you have experience doing a project so that the science I think you have to present and the best part of these and the thing that you have to keep in mind is that you have to believe in yourself. If you believe in yourself, when you're trying to present to an interview, they are going to really find you. So you first have to believe that you are either the scientist and then they are going to see this on you.

[00:27:36] Confidence is so challenging in the interview because people that land one Data science offer, they will quickly get two or three more because they have so much confidence and people that are struggling to get the first offer that comes up in the interview and no one wants. It's like dating. If no one wants you, no one wants you. If everyone wants you, everyone wants you. And so people that interview a lot and get good at it, they have the confidence, especially if you if you can get a job offer as an analyst here. So confidence is it's the maker or breaker for people in the interview process.

[00:28:06] And you don't even need to have actual job interviews lined up to get interview experience, like do mock interviews. I go on to any one of my posts and see people commenting on their or liking it better, aspiring Data scientist, whatever. It hit them up and say, hey, can you hook me up with a mock interview, reach out to people in the community and LinkedIn and just give mock interviews to every day. Go, guys, sit up.

[00:28:30] Carlos, from interviews you'll regret saying that. No, no, I don't write my reviews there. I'm people. I'm doing it my way. I look at it, I promise. Do you want to kind of attack with the people who finished?

[00:28:45] Projects are more valuable than people who have a lot of ideas was like, oh yeah. Saying I can't finish anything that I start for some reason. But also I just want to know, like with taking beginner and stuff off your resume and out of your language, it's really important to do that because I'll get resumes from people and they'll say like fifty technologies. And then I'm like, OK, you don't have the technology and then I'll see resumes or it's just like Python Intermediate are beginner. And I'm like, dude, like I'd rather be the first guy who claims to know everything honestly. Like you're like, don't sell yourself short and you're not expected to know everything. Honestly, if it's on Google it's like an inch away from your brain anyway.

[00:29:23] So just just to Googling, I've seen so many resumes with amnesty and the Twitter sentiment example, like like literally like one on one hello. World Amnesty Plan or CNN's Twitter.

[00:29:35] Like Backwords sentiment, chronic disease for Titanic, like throw in the IRS, Data said people.

[00:29:43] Oh, come on now, people don't be insulting Titanic. That's like one of my favorites. Yeah.

[00:29:47] You know, it wasn't that useful for Zyprexa. That's useful for practicing. But it shouldn't be the star of your resume because you want to be like you want to have a niche that you're chasing, like, oh, I want to do public health. So this my public health focused. I want to finance my finance focused project, I see resumes. The project is deep learning for finance and then cancer detection, and they want to do like retail analytics. I'm just like, you're not you're not pointed in the same direction.

[00:30:17] You got to be consistent. At least have a focus and pursue that focus on open it up to see if either Monáe or Toshie or Sooraj or Ashanti's questions. I think, man, I think you messaged me on LinkedIn saying that you were really looking for topsiders. I'll give you the floor first.

[00:30:36] Well, I did message you, but I don't currently have any questions. Yeah, just listening right now.

[00:30:43] Right on. Right. I will enjoy the the chatter. So the technical question in the chat. OK, yeah, definitely. Eric, if you have a question, go ahead. I don't see you on mute yourself and know that you're in the Middle East yourself and go for it.

[00:30:56] Yeah. I was wondering, I have seen unit testing come up all over the place. I think I think I've definitely seen it, Carlos, and a few of your posts saying do it, but I don't actually even know what unit testing is. And I was hoping you could explain that.

[00:31:09] So making sure that big things play up, like how much of Data science should it be? Software engineering, because people will just go to Harp wondering what are they doing other doing agile dev setups, what OK, Data science team do actual text box, whatever, human centered design, whatever. So that's just the so unit testing is the idea that if you have an input and you have a desired output, that input should always have that output, no matter what setting and what context. And you do the test in terms of like, oh, give my input as class character. I always get this return or this error. So it's just having a full plan for all of you. It can be functional. It could be any module really getting the platform level stuff. That's I think that's one where Thomas and Dave's skill set, you have different kinds of unit tests, but the main idea is inputs and outputs that always match no matter what, given some set of instructions.

[00:31:59] So basically it's trying to break whatever you built in as many ways as possible.

[00:32:04] Stepping for that could also save you from breaking with like a version of grade, because the deep learning frameworks are infamous for this. That aversion upgrade will break all your old models or even change the outputs. So it saves you from that.

[00:32:17] So, Eric, you're a sure, I believe, right?

[00:32:21] I'm more comfortable in Python, but yeah.

[00:32:23] Comisar so Python, you can you can explore the API test package and get familiar with that. That really helps automate test design type stuff.

[00:32:32] So ah, there's a test that Data better.

[00:32:36] And I would and I would jump all over Carloss saying if you're looking for an analytics position and they're asking you about unit testing, you need to ask why, why? My value proposition is not writing necessarily very high quality software. My my value proposition is analyzing the data and driving insights. Do you really need to test for that? Who don't? So be sure to ask that further. Part of the earlier question that we were talking about, I think that was before you joined IRC.

[00:33:04] So where I go, I do write unit tests. So I'm actually curious the like for like how you distinguish between like and you write the unit test with like software engineering software engineering component. That's like implementing a Data science component versus like writing a unit test for like my analytics is. That makes sense.

[00:33:24] So here's the rule of thumb. I is am I going to get called in the middle of the night? If it blows up? If the answer is no, you don't need a unit test for it. That's a good rule of thumb. I like that. If that's because I used to get called in the middle of the night.

[00:33:35] But it's not technical debt if the debt doesn't come due is dangerous. But the dangerous one. I like it. I like that.

[00:33:43] I have just a slight counter to all of this started and I don't disagree with what was said. But if you're in a role where you're in a frequent release production environment, I was doing algorithm development is pretty much Data scientific kind of that and a little different, a little more. We absolutely had to rely on a massive system before release to make sure that our changes didn't mess anything up. And it's a lot like what's been said. But once the system is built, then you've got really good research statements where you absolutely can't have a failure. It's really not too hard to build up and maintain its system. But Dave's point well taken, too, depending on your role, it may be crazy for you to be the one expected to write fast and maintain them.

[00:34:30] So let's turn it over to Tashi. She has a question. And if she doesn't have a question, then we'll go to to either Sooraj or in the shop.

[00:34:39] Yeah, I don't really have a question right now. So we're loving it. They're definitely asking me, well, I'm happy you're here. You're not supposed to have a question. No, nothing is awesome. And I've got a question from Afghanistana in the chat to I should I do that?

[00:34:55] There's so many people in Washington is either Austin has a God given gift of first of all.

[00:35:00] Yeah. Good discussion so far. I wanted to. Of the salary, not like think, because my after my first my first job as a contractor was very low paying. So like when a recruiter reached out to me for a different job, I was just like, yeah, just whatever. He asked me straight forward, like, what are you making right now, if you don't mind telling us? Yeah, I'm just making the market whatever the market stands for this position. And I was like a significant bump into my new role. So that's that's that was a really nice thing. And I like lesson I learned. And also, like everyone, I don't know what everyone's experience been with the recruiters, but for for my experience, they're like not in a bad way, but they're just not non-technical. So I could get away with, you know, like like like talking about not getting into the weeds with technical stuff, just letting them know, hey, this is how I am as a person. If you think is a good fit, you know, let me know. Like, basically when my question was from this post I saw the other day that Jason Crans posted on LinkedIn basically hotsy I. Yeah. So yeah, he just he was talking about it's going to be a shortage of business domain experts and the Data science and Alex face. So I just wanted to see your guys opinion on it, like what's your observation or predictions if you have any.

[00:36:23] He said there's going to be a shortage of business domain experts in the Data science space we have.

[00:36:29] We have a significant shortage of business domain experts in the Data, AIs and space.

[00:36:34] That's what I think. That's because people are just learning all the tools but not figuring out how to apply them. So going back to Carloss Point, it's OK, quit doing the Titanic project and, you know, Martz's classification or whatever, and Nesh down. Right? I understand that there are different industries. Do some case studies learn what problems those Data scientists in the industry is facing and then try to replicate those problems? I see Garren has some input here. Go for it.

[00:37:04] A Yeah, I had actually a question about, you know, that transition from a data analyst position to, let's say, considering I want to move towards a data scientist position and considering, you know, and that that was the question for the larger group was, you know, of course, a lot of what I might be able to explain would be, you know, my personal projects, everything that I've done apart from my work, because I don't use, like a lot of statistics unsterile. So it's a lot of Tablo and other tools of basic knowledge tools. So I want to understand, like from an individual perspective, is is it in any way going to affect you know, it has to be, for example, different data scientists who have been in that position for versus somebody who is trying to switch that switch to that position. And is that something, you know, maybe theoretically it might be possible, but I wanted to get you know, because we have everybody here who have a lot of experience. Is that is that something that works out? Yeah. How do we look at it? Yeah.

[00:38:18] So let's do this let's first get this question answered and then let's get to the core of what your question was that there's a lot going on there. So to address this question, does either Ben, Carlos, Tom, Dave, Giovana, Kate, anybody want to chime in and say, is there really I wonder if it's just kind of writing the theme or that way?

[00:38:40] Because if I if I'm speaking, I say Data scientists struggle with communication and business acumen and everyone's like, yeah, we've known that for a few years now. So I wonder if it's just a continuation of that where my quick answer to this is I don't I think Data scientists need to start. They need to stop trying to own the the business acumen in the domain. Just get the people in the room who are the subject matter experts, who are the people that care, who are the people that can actually validate the results, like move way past a statistical validation score. What's the KPI? What's the dollar amount you're saving? If that person is not at the meetings at the beginning, in the end, then you have a problem. And I don't see that as a Data science problem. It's more just it's more networking and working well within the organization.

[00:39:22] Yeah, I would I would wonder to the point of ations question and by the way, I tend to agree with most of the things that Jason Krantz transpose, by the way. And if you see my responses to his posts, you know that I would argue that if you think of Data science as being purely an exercise in technical virtuosity, do I know deep learning? Can I production quality code? Can I do this? Can I do that? It's arguably not significantly different than just typical software engineering roles in general. And there's been a long group of business people with software engineers since way back in the day. If you're interested in using data to drive business results, call that a Data analyst. Call a data scientist. I don't care to. Really matter if you want to use Data to drive business results, of course, it helps if you know the business course. And in fact, I would argue that if you want to make the most money, if you want to make the most money in a big corporation, you don't worry about technical virtuosity. You worry about business knowledge. And you'd be surprised how little statistics and how little machine learning and how little of this and that you need to know to actually make business impact and be an adviser to the executives. And that's where the real money comes in, generally speaking. So just to say what you want to do is geeking out is what you're passionate about. Sweet, become an engineer, whatever grade. Do it. If you want to do analytics. Yeah, you need to you need to have business knowledge to be most effective, in my experience.

[00:40:44] And to field this question with Karen's question, I think if you're currently denounced by the Tablo and you don't know what to do to make the next step, figure out, are people actually looking at my Tablo outputs, what are they doing with this tableau out?

[00:40:58] But what would it look like for me to just throw in a forecast and outputs like make any steps toward other people asking about modeling and Data Stintz work ethic, except that it's way easier to just like get a few people hooked on acne for Data fine stuff and free to switch to a totally new company and have them as of science test. I mean, if you're already doing good stuff with Data and SQL and top like probably halfway there to something.

[00:41:24] So Monaca, do you have any comments on that?

[00:41:27] Yeah, I mean, at the end of the day, Data analyst, data scientist, the main point that you're trying to do is to solve a problem, to help the business be better, make more money, get more customers.

[00:41:41] You're just trying to solve a problem. So any time that you are speaking to those stakeholders and you're presenting your results, you want to kind of start with the fact that you solve the problem and then not go into, as previously stated, not go into the deep dove into what tools were used, because the high level, the high level people don't really care. But why you how you solved it and how that's going to make their lives or their business better and easier.

[00:42:11] Yeah, I mean, I like to think as a data scientist, as methodologies, it's like at the core of what we do is we use methodology to solve problems and methodology. I mean, does it rely on tools, maybe to a certain extent, but not necessarily the only thing that we use. So Robbie joined in. That is awesome. Nashat you've got a question if you do go for it and everybody help me welcoming Sarah as well. Good to have you here. Sarah. Hi. Thank you.

[00:42:43] Hi guys. First of all, I'm very happy to see all the inspiring personalities that I've been following on LinkedIn is the first time I'm getting a chance to meet you and talk with you. So it is a great opportunity to thank you, Patrick, for setting this up here for the first time. So my one question. If you guys are being working, contributing to the community, having family, so you guys have a lot of responsibility how you guys manage your time. I know many people have been talking on this topic, but I. I want to hear more about that. Can you give some input on how you manage your time? You work on specific based on your content. I just want to show us some input on that.

[00:43:40] And definitely let's hear from days go for it. I was going to make a bad joke. So these days it's sleeping at two, three o'clock in the morning after studying and then having like six of the coffees during the summer time management out the window. So sleep slows everything. But there's there's a lot to do and a lot to learn. So you just gotta to figure out what's your balance that you can do before you go crazy.

[00:44:00] It's a balance. And clearly, I love to hear from either Sarah or Jeevana. How do you guys manage time?

[00:44:06] So if I can step in here. Well, I know in the midst of all of covid and whatnot, I'll speak from my own experience is that for a while I had a hard time doing that. So I had to think about take a step back from kind of all the responsibilities and all the things that were getting thrown at me and kind of say no to things that I would have otherwise, in a normal situation, been, you know, being able to handle. And so taking a step back, thinking about myself and accommodating what I needed in this moment, especially with, you know, kind of everything that's going on. And then once you diagnose and are able to accommodate yourself in this moment, kind of to what Haseeb said, just like a lot of coffee and then and also knowing how to say no to things and yes to the things that are appropriate. That makes sense, and so something that I've actually wanted to start doing recently is coming up with what my one three year plans are, because I have a lot of different initiatives that I want to take on. But understanding where I kind of want to lead into and where I want to focus, my energy is important.

[00:45:18] And so knowing what that is for yourself and taking on the opportunities that allow you to get closer to where your goals are is important.

[00:45:28] Yeah, I would like to that because. Excellent. The answer. I love that point of view is that I think we all of us, we work on priorities.

[00:45:41] Maybe we want to do a lot of things during the week. But I think we have priorities and I think we have to focus on that because maybe I want to do a lot of projects. I want to help mentoring a lot of people. And but I, I need to continue writing in the things that I want to develop as a professional in my personal life. But I think when you have a priority, it helps you to guide you to achieve your goals and your personal statement, because help the community be part of the community or with the community, grow with the community are the priorities. So I think when you have the passion, as you said before, you drink more coffee because you know that this is not just for you, is for a lot of people that is trying to grow. And when you were with other people, is that you connect with that energy and I think you don't think about that. You don't have energy or you don't have time. I think you are going to do this because a lot of people is supporting you. And I know you want to think some because she's they like you every day, inspire people and they think, is that the way that he proposed Ibbs is? It's a way of giving me personally energy to continue thinking how how can I help others and at the same time helping myself. I think this is a source of inspiration.

[00:47:21] I really love that great perspectives. And I mean just for me and whoever else wants to chime in after me, please go for it. I'm having a full time job. Davis AIs dream job, mentorship platform, podcast and a brand new baby like the plate is full.

[00:47:35] So I only do things that are going to move me towards my goals. For a while I was like, Oh, I'm going to write blog posts and then will writing blog posts. As wonderful as that is, as much as I want to do that, it's not going to move me towards where I'm wanting to be. Right. So I just get clear on precisely what it is that I want to achieve, what I want to accomplish. And once I get clarity on that, I will only do things that support that objective. So this requires introspection like I did. I spent last Saturday. I sat like it was like six a.m. I sat by myself for like two hours and just. No, no, no pain, no paper, no distractions, just a thought. Not real clear on exactly what it was that I wanted to accomplish by March of next year. And now everything that I do only supports those objectives. So clarity is key. You need to know exactly what it is that you want to accomplish. Monica, Ben, anybody else would chime in with some time management tips.

[00:48:36] So I'm a big fan of sucking at your weaknesses. So don't try to fix your weaknesses. Just put a Band-Aid over them and focus on what your superpowers are. So try to like, really focus like what is my superpower? What where am I better than my peers? And dump all your energy into that because you can't be good at everything. And then the other thing that I think is helpful for people that post content on LinkedIn, the thing that's been helpful for me is during the day, if I think of something, if I even on this call, like I just posted something, someone says something like, oh, this is interesting, you post it that way. I'm not stuck staring at a computer screen. Scratch my head in the morning, wasting time about what would I write in the best post. And here's what other people think. My best posts are the ones where I'm being emotional or I'm pissed or I'm venting about something that maybe I should really delete the post. And those are the ones that get the most traction.

[00:49:27] And one thing I do, because there's lots of ideas that that float around, right. When you think about stuff to do is have a period of time where you just write down all of your ideas just to kind of get it out of your head into the physical world. Right. And then kind of organize them by this is probably one party, too, so and so forth. And I typically plan my next day, the night before, and I'll take I have, like, no cards right here and and I'll write down the date. And just like three things I want to accomplish. And it's never more than two or three because realistically, that's, you know. That's enough to keep you busy the entire day. So, yeah, I'm going to open it up to anybody else.

[00:50:07] Yeah, I'm going to be the first really quick. I would say just listen to your body. I did this whole thing in undergrad about the embodiment of mind and how, like, you're just you're not your brain like you are an awareness of like a whole consciousness through your whole body and stuff. And that is like if you're not feeling like doing something, just acknowledge like, oh, I can see Data for two hours staring at the side or I can go take a walk and like go read a book and do something else, like just listen, like leverage your flow state, like if you're like super productive in the zone as long as you can, but you're not in the zone. Like don't just sit there and waste time beating yourself up over not doing stuff. Like my mentees will waste two weeks being mad that they didn't get me analysis that I asked for in time. And I'm like, wait, you wasted two weeks because you didn't finish something. Like I stole ten dollars from your bank account, like, doesn't make any sense. So like, listen to yourself. I don't plan on my desk. That's my big no, I don't play anything. I'm just goofing around.

[00:51:01] So I'm a big fan of like simple rules like just like either almost even cliche. So like one. And by the way, I think this is everything I'm about ready to say is actually more than just about whether you want to be an influence or on LinkedIn or whatever. It's also about if you want to work in applied technology in the business environment. I think all three of these things are exactly applicable to both sides. One, it's not a job, it's a lifestyle choice. My background, software engineering, that's how I started back in the day. I was a software engineer. It's not a job. It's a lifestyle choice. If you want to stay abreast of latest greatest, you're going to have to invest time after work nights and weekends forever. That's just the way it is. You're just choosing that lifestyle to relentless prioritization or position each and every day, every hour of every day, week in, week out, and then three. And this one I learned from a gym at Microsoft. If it's not on a list, it doesn't exist. So make lists. Bekerman No pad is just some sort of app. We're just writing down a paper. Those three things typically are how I manage what I do and say, look, you know what, it's long hours, just part of the deal, OK? So it's all about relentlessly prioritizing. I spend that time basically.

[00:52:12] So those three things, you're probably pretty good if you follow them every day. So if I can summarize everything that everyone has said, it's a combination of compassion towards yourself. Just if you don't make progress on your goals, that is OK. Just to get right back on track, capturing what it is that you want to get done, get that out of your head into the physical world, prioritize it, use a list and then track your progress against that, whether it's daily, weekly, hourly, if you're that neurotic. But I think that kind of summarizes the advice that everybody has has given here. Let's move on.

[00:52:47] You hear that? He said you're neurotic.

[00:52:48] I'm just kidding. Yeah, I picked up on that. Thanks, man. Appreciate it, dude.

[00:52:55] Dude, I'm the same way I was following the Brendon Bouchard's high performance calendar method for a while. And as I do, I can't fucking do this. I can't be this granular with my day. So now it's just here's the three things I want to accomplish today and just focus towards that. Let's see if we got any questions from anybody. And there's so many people on screen, so current. So you had your hand up. You got a question. Definitely go for it. If not, then shot that shot this calling from India. So it's like 4:00 a.m. for her. So if you got a question, you should go for it.

[00:53:28] I don't know any question. I just had to listen to everybody.

[00:53:34] Thanks for joining. I have a question that no one else does. I want go for it. Yes. I'm on the ground floor of a new client.

[00:53:41] They have they have no formal data collection, no data storage, no real plans. All they know is that this data is fundamentally spatial and they need network analysis done. And I have no clue what to give me. That will make sense to them because I know a bunch of miscellaneous things that I know works, but I don't know what will actually advance their stories for them. I mean, I'm not going to calculate a bunch of, again, centrality and then be like, here you go. Let's put this like I could do all kinds of stuff that won't work for them. And I don't know what I can't really figure out. What's the simple first step? What's the Data type? Well, it's it's it's like under an NDA, it's like international migration patterns for different disease outbreaks. So they give a developing world like the Data very narrative heavy. It's not formalized in any way. And they're just hoping to find some patterns or clusters or stories out of that Data they think there's something to do with network somewhere and they just want me to find it if it exists and tell them what it is.

[00:54:37] Ok, so this isn't a perfect fit for you, but it's a funny thing of experience. If you show any type of clustering or really advanced machine learning that begins to form some type of structure, everyone on the call will celebrate like we're so excited executives won't react.

[00:54:53] As soon as you color coded with a KPI, they're on the edge of their seat. They need to know what is happening inside this group or. Inside this cluster, inside this visualizer, so whatever type of analysis you do color, cut it with KPI that they know and they'll be paying very close attention to anything that's polarizing. That's helpful. I've done it the wrong way plenty of times.

[00:55:15] Harp, can I just type in on one central theme I want to make? There's brilliance in this group and I want to make sure that a brilliant thing that's being said round and round is not lost. It comes back to something. Everyone's touched on it a little bit. But sometimes what we're taking to our internal partners can seem like salesmanship. But I think what we're really touching on, it gets more to, of course, like being just illustrated, understanding what really lights their fire. But probably more importantly, just it's more like relationship building, you know, really selling someone on. I want a business relationship with you. I want to bring the value of Data to you and giving it to them in small enough pieces in their language that they speak like them just illustrated. If we just do that, that's key.

[00:56:11] Learn to build, learn the self you do both, you'll be unstoppable as a Data scientist for sure. Joe, my man is so good to see you here.

[00:56:19] This is good to see you too. I think I got the time wrong, but it's not always men will still go.

[00:56:25] I got I got nothing to do after this.

[00:56:28] I'll show up next week. But every week are like, what's the how does this work?

[00:56:32] Yeah, I hold them every week. I can't promise I'll always need this stacked every single week, but I can't promise at the very least it'll be me here so that that is my promise. So let's see any questions from either a shit or I think Soueid said he did not have a question. What about where did Nicholas go? I thought I was like, oh, there he is. Nicholas, are you on the spot here?

[00:56:55] Don't have any pressing questions right now. I like I like the talk about unit testing. I mean, I would think I was just doing a case study the other day two days ago, and I found myself type in the keyword, a certain python and realizing I have no idea what it did, but it sounded like the right thing to do. OK, study. And then I quickly stopped it and took it out and didn't know if I should be including that in like a simple Jupiter notebook model building case study. But it seemed wrong. So I ended up not doing it. And I think that's the right choice, given what we talked about production and if there's not a one a.m. call, which there wouldn't be, and something that's simple, I should've included a degree.

[00:57:35] Yeah. One one resource that I really enjoyed that helped me kind of understand. Like, I mean, I don't do hard core testing with my code, but something that really helped me was I think his name is Ted Patro, Dunder Data. He had, of course, that was building a data and analytics or data analysis library from scratch, and that was done heavily from the test driven development aspect of things. And that kind of just gave me an appreciation and awareness for how that fits into the data science lifecycle.

[00:58:02] So definitely it looks like ten bucks or something like that is super cheap, but well worth it when you type his name and of course, are the module name in the comments.

[00:58:11] Yeah, I'll definitely do that. Yeah. So I just I just feel obligated to say something here and yeah, I might lose all if I had any respect from anybody here, I just lose it all. It's OK. I'm going to go. I'm going to go. I'm going to go for it. Anyway, I used to be the guy software architect the team lead where we would have a hat and if anybody broke the built because of test failure, they had to wear essentially the dunce hat. So I'm just using that to establish a baseline for about what I'm from for what I'm about ready to say. I have not wrote a unit test in years, years, and the reason for that is I analyzed data. I don't build production software systems, analyze data. So I know about unit tests, I know about software engineering. And I also know when it's not actually needed and when it's overkill and it's not effective for the business from a cost and time perspective. So I would just advise everyone just to think of that in those terms. Software engineering is not free. It costs money, it takes time. Is it really worth it? Just think about that.

[00:59:11] Although I would say to add to that, not in a way like some of my friends work on this new package. It's a great expectations. So it's sort of unit testing your Data for Data quality.

[00:59:26] So we find ourselves more and more, I think, suggesting that people use this as just a way to sanity.

[00:59:31] Check the property for Data. So might be something worth checking up at today's point. There's a certain threshold. I'd say you're not writing software, analyzing data. Right. So but that said, I think there are tools out there that allow you to test your data now. So that's pretty cool.

[00:59:48] Would you say if I was about to say if I went on, I was if I was in Joe's world with this Data engineer added tax writing my DB unit tests just like I used to do in the old days. Right. But luckily, I'm an aunt and I'm an analytics professional. I got people.

[01:00:02] You got two people handling this. Forget it. Just do your job.

[01:00:05] So why waste brain cycles, Dave, if you're learning better, something else, like if you're really good analyzing Data get better at analyzing data and don't waste hours and hours and hours of time writing tests like that. So the thing that people don't really talk about that much openly is we are all dying like literally like we have two billion heartbeats in a day and some of them are coming sooner. Every breath we take is closer to the end. And so spend time working on things you like. Don't write unit tests if you don't need them. We're smarter than that anyway.

[01:00:38] No, actually that that like that to go off the deep end to really say comtemplate mortality every single day.

[01:00:44] Remind myself I like translating failed models into life lost. So yes, you died a little bit with that failed model.

[01:00:53] I was going to ask the question, would you say that unit testing, test driven development is more important for Data engineers than it is for data scientist? But going based on what they were saying when he's talking Data, it seems like that that's the case, right?

[01:01:07] I think it depends what you're doing. And Data engineering, there's sort of a continuum with Data engineering where on one spectrum you're possibly writing a lot of code or using open source libraries and appropriating those. So it's a code heavy version and I would definitely suggest writing unit tests or regression tests or any sort of tests to make sure your system works. Not doing that. I would say is professionally negligent. You should get fired for doing that. That said, there's a lot of managed services out there where engineers on teams just going to do this for you and you can pay them to do this. And the way I see things going now, you know, we just record a video on this on Monday, but know sort of the continuum of really writing public versus using managed open source versus using third party tools. I think you can get away from it like Tego Bence Grim Reaper points. If it's something that's not in your core competency, why don't you just pay somebody else to do it? Like the argument I always use is you. What do you need to get new tires for your car? Like you just make yours from scratch, like you've got to download the plans off the Internet and like go to the tire parts store and just like nail them and make. I don't think anyone does that in a screw. Right. So why would you do that with your things in your core competency? So but that said, if you have to write code, right. Unicast, just suppose that for engineering related stuff. So if you don't, then I guess you don't know it's an investment in your time so you can invest the time up front now to avoid, I guess, alarming things happening to you later.

[01:02:39] So so let me make this up to either Monica, Giovana or Sara.

[01:02:44] What are some things that data scientist are not good at, but they probably should start getting good at of communication as I came in pause forever and just communicating with the right individuals as well.

[01:03:08] Lots of times I'll spend hours and hours trying to search for the right person and then I'll have a conversation with that person. And then at the end they're like, Oh yeah, that's not even me. And so getting everybody in the right room and also being able to communicate between the business people and the technology people kind of being that middle man and knowing the different languages and how to talk tech, talk with the tech people. But you can't transfer that business people. They won't know what you're talking about. And so if you know the two different languages, you'll be able to get what you need from each.

[01:03:49] Yeah, I totally agree with Monica. And I think you have to think when you prepare your visa applications, all the things that you have to show to your stakeholders or your team, you have to think in your audience with the people that Monica discovered. That's another thing that someone mentioned here and I think has been is to build a network inside the company. Because I said at the time, if you need information from a lot of the themes, for example, sometimes if you get the data and you have problems, though, because the data is not in a proper way, what we know that this is common, but maybe talking with a person who is going to enter or prepare the data that is going to to be part of your dataset, I think maybe you can gain a lot if you talk with a team that has to enter the data into that asset or if they prepare, for example, the forms that the customer has to. To feel to have the information, I think, connect with all these things, maybe it is not just to send an e-mail. Sometimes you have to go to the desk to meet these people in person, have a relationship with these people. And they when you send an email, they may said she's going to be different because they know you and you know them and you know how to talk with them. And if they have time, what time they have, they can answer your emails. You can connect better if you communicate better. Another thing is the thing work. So you have to think that you are not alone. You are not building your your mother alone. You do need you need a lot of people, a lot of information to be these more than properly, because even more that is going to give an output that is helpful for everyone in the company.

[01:05:59] I absolutely love that. And I think another important point is the ability to ask questions and find problems. I think, Sarah, I'd love for you to speak on this because I absolutely love your headline on LinkedIn words question. Everything ends with Data. Do you think do scientists need to develop their question asking abilities? And if so, how do they do that 100 percent?

[01:06:21] One more thing that I wanted to add to it, Monica and Ivana said, was that sometimes the companies are it's it makes it that. So what they're saying is basically that communication or relationships is a big part of your success as a data scientist, Data analyst wherever you sit in the company. But sometimes a company by design or by construction is created in a way where there is isolation and it's a lot harder to break the barriers and get to the people who are your stakeholders. So there's a lot more effort that needs to be made, I think, from depending on where you sit in the company to forge those relationships. And so, like, for example, I work on the revenue strategy team and I'm very connected to our sales team. So the models that I end up producing versus the, you know, the engineering or data science team that is a little bit further in relation to the sales team end up making naive assumptions, potentially because their involvement and their kind of relationships are a little bit more spread out. So depending on where you sit, you may have to put in a little bit extra time and effort to create those relationships.

[01:07:33] In terms of curiosity. Yes, I think this is huge. I think one thing we don't do well enough that I try to encourage and even when I was instructing, is that you can learn the tools and the tools are a means to kind of getting the result that you want. But if you don't know what you're aiming for and you don't know what questions you're asking and you're not helping guide the conversations in the right way, it ends up being a lot of unfortunately wasted time potentially on things where the KPI is a success. Metrics weren't defined up front, and so to someone in the room is needed to kind of align cross-functional partners to to kind of where the goals are and to asking the right questions and then and then in the back end, kind of getting the right data set. I think that Kathy, who did a really good talk on kind of inherited Data versus Data that you've constructed yourself and so kind of understanding what types of Data you're dealing with and whether or not you need to go seek that yourself as a result of the new questions or the guidance that you're getting.

[01:08:40] So when it comes to KPIs and metrics and stuff like that, that's really not something that I think you can get when you're just going to keep boot camp or doing personal projects. Right. So how do we develop an understanding and awareness and appreciation for constructing KPIs, understanding how they impact the business? Is there to read reports from from companies. Look at the financial reports we like. Is that question making sense?

[01:09:09] Yeah, I want to say a lot of this. I mean, probably other people have have opinions on this as well as I read a lot, especially in the domain that I work. And from that I just generate conversations whether or not people are initiating them or not. I just I just start creating conversation where I feel like maybe the the response is just like this group. Right. Like we all value each other's opinions. How can we just put questions out there, things that are on our mind and just having a discussion about it. And so so I would just say continue to read up even when you don't feel like reading something, if it pops up in your email, just open it, open it, see if it's interesting and then maybe try having a conversation with people about it, something really quick.

[01:09:55] So I get a lot of emails from potential vendors and they don't do their research, I don't know, or consulting firm that we're not. We're not. And for them anyway. But what I'll do is I'll actually ask them, can you give me a case study of what you've done for others and what I've got? And I was like crazy, useful case study that I've been able to use in my business development. I'll do informational interviews with their team, learn about their products. And then I'm like, oh, like I have a best friend at my box. I have a best friend that like all these other vendors, I've Data about whatever on Data robot because Michael left. But anyway, so yeah. I mean like it's really interesting how much information people just give you for free that like super usable. So like follow those emails, meet random people and random industries like networking. So useful. I think that is relevant. Sorry I wasn't.

[01:10:41] No, definitely. Definitely relevant. So open it up for maybe like the last question here. And if not, then we'll wrap it up, guys. Thank you so much for everybody that came out. Last minute questions. Now is the chance we're going to put the chat notes.

[01:10:56] Yeah, yeah.

[01:10:57] Definitely chat notes. They're going to be saved in the show notes both on YouTube and on the podcast thing. I'll delineate them with a bunch of asterisks so people can read them. I mean, thank you so much, everybody. I showed up. I was not expecting like the entire LinkedIn community of Data science influence to show up at one place at one time. That's mind boggling. But thank you so much for coming through. I hope you guys show up again in the future. We've got officers again next week, even though it is Black Friday. I live in Canada, so whatever. It's just regular Friday. And then we've got like a few more of these officers left for the year. Before we take a break, I'm looking at the calendar. Last officer before break, even be on December 18th. And we'll take a break for a couple of weeks, obviously, for Christmas and New Year's. Thank you so much, everybody, for showing up. Hope you come back. Super happy to have you guys here. Take care. Have a good rest of the weekend and we will see you around.

[01:11:53] Thanks, everyone. Things are great. Thank you so much for putting this on. Thank you. Thank you.

[01:12:00] Bye bye, everyone.