comet-ml-oh-may30.mp3

Speaker1: [00:00:09] What's up, everybody? Welcome to the comet and our office hours. It is Sunday, May 30th, a long weekend in the States, but I'm still happy that you guys have been able to make it here and hang out with us today. We are powered by the artists of Data Science, so I'm excited to be here. And this is our session with the guys. We got some good friends in the chat. My friend Kevin Kevin's in building. Austin is here. So is Cristoff. Man, good to see you. Guys here are excited to get this session started, man. So how's everybody's weekend?

Speaker2: [00:00:43] Trying to figure out who the rule was to put a session on Sunday on Memorial Day weekend propranolol in Canada.

Speaker1: [00:00:50] So it doesn't really matter to me what what it is. And plus, if you're trying to get somewhere in life, it doesn't matter if it's a holiday or not. You show up and you ask the questions and you get the information you need to move ahead. That being said,

Speaker2: [00:01:01] I mean, it doesn't really matter

Speaker1: [00:01:03] Me. It's a holiday or not. I still out here. I'm still grounded. I'm still grounded now. And plus, there's a lot of people in Europe and I'm pretty sure they don't celebrate Memorial Day in Europe. So there's that as well. It's been incredibly busy at work and with everything else. I had an opportunity to interview several people this week for the podcast, including Liz FastLane, who wrote the book No Hard Feelings. You might recognize Liz from the work she does on Instagram under the handle. Liz and Molly, she does this really cool illustrations and things like that. I got a chance to chat with her. I spoke to Jordan Ellenberg, who wrote How Not to Be Wrong Power of Mathematical Thinking, as well as shape as a really interesting conversation. And then one of my good friends, Argin, such they've got a chance to interview him for my podcast for one of my conversation series episodes. I was talking to my buddy about this yesterday. Like, believe it or not, I don't have many friends, just a few. And even those few that I do have, I don't get a chance to speak with them as much. But it dawned on me yesterday that pretty much every conversation I have is recorded and made public in some way, shape or form, whether they're the the conversations I'm having at work for this discovery phase I'm doing for a project, all those conversations get recorded, whether it is the office hours I do as part of a Data science dream job. All that stuff is recorded. The podcast, all those conversations, recorded office, everything is recorded. And it just it made me feel a little a little strange, but it keeps me honest, I guess. So you guys, you know, I'm not I'm not bullshitting. I'm in I'm in the streets out here with this Data some stuff. Man, I

Speaker2: [00:02:41] Got a question for you. How does that make you, like, want to actively change that a little bit? I mean, like coming out of like, you know, sort of coming out of the quote unquote pandemic and having more opportunities to see people and talk to people in different ways. Do you feel like you have a way to come out of that? A little bit.

Speaker1: [00:02:55] I mean, I'm hoping to to do more types of events where maybe we're doing them like, you know, in person life type of events. But I know there's going to be hard given the the state of the world. But I mean, I do miss I do miss being around people. That is that is for sure, man, because I'm usually just around my wife and my baby, even though I love them very much. Um, I enjoy talking to everybody that I talked to through to these officers and stuff. But I mean, it's it's tough when you don't have time for many close relationships because you know that all of my time is spent doing things right. I don't know, just going off on a non related topic there. We could we could bring it back to today's AIs if you want. But if anybody else is feeling this type of Zoome fatigue, this type of fatigue, where you just feel like you're always having to interact with people on screen and not in person, like I feel you like I'm, you know, dealing with those same struggles. What about you, man? What's it like for you?

Speaker2: [00:03:54] Yeah, I mean, I so this was my third week, a comment as the head of community. So, you know, I've never met any of my team actually in person. We're all around the world as we have an office in Israel that we just opened in New York City. And they're thinking about sort of expanding out west. And I'm in the Boston area of the United States. So I'm looking forward to taking a train ride into New York and meet some folks. But the onboarding remotely is a particular kind of zone fatigue, where on the one hand I come in like with a lot of energy and I want to start really off well and make a good impression and sort of jump in and start, you know, hit the ground running because I feel like I have a lot to there's a lot to do and a lot to a lot to improve on what we do. But then also sort of like trying to figure out where I fit in to different conversations. And that has its own sort of mental fatigue. That's that's definitely related to Zuman being distance and things like that. So overall, it's going really well, but I'm having a lot of fun with that. But it is yeah, there's a certain kind of like deflation almost of like starting a new job and not getting to be in the office and people can be around. So I'm excited to, like, make it to New York and at least, you know, hey, these people are real and they're not just like on the other end of the screen for sure.

Speaker1: [00:05:00] Yeah. That's something that I was talking to Liz FastLane about. We brought that up in our discussion for the podcast. And she was like, you know, onboarding into a. Company during this pandemic era where everything is remote, but you're joining teams where people have formed those relationships one on one already, they've gone out to lunch to get coffee together. They take breaks together and you don't get a chance to form. Those types of relationships can be tough. And now we've got it. It's the responsibility falls on kind of both parties to make sure if you have a new team member coming into your team that's working remotely, you know, go out of your way to make that person feel welcome and let them know why you're excited for them to join the team. And on the flip side, if you're joining a new new company to find people to go message and chat with and do virtual coffee dates and stuff works because at the end of the day, man, we might be doing Data science and machine learning, but that doesn't mean that we automate away the human aspect of the work we do. I think that's the most important part of what we do. Yeah, I'm excited to take some questions, so if anybody has got questions, go ahead and let me know. I see some new names here. Aren't these in the building? Aren't you? How's it going, man? We got a project about Cristoff A.M.. How's it going, man?

Speaker2: [00:06:20] Did not no running today. Just the outpost. Jim. Yeah.

Speaker1: [00:06:26] Talk to us about exercise. Just about this interesting project. So you posted a really cool project about Disney movies, and it was like this Tableau Dashboard. Talk to us about what they are more interested in, how you decided to pursue that particular topic, because I think that's something that we don't really get taught. How do we make it? How to use particular tools? We make it taught how to use particular methodologies and and particular tools, but we're not really taught how to come up with the interesting problem to work on. And I think that's probably the number one important thing that you should do or focus on, rather, when you're making a portfolio project is what is it that I am going to actually work on. So how do you how do you come up with that that particular topic of Disney movies?

Speaker2: [00:07:16] Quite enough. It was, I would say by chance, this one. And it's not a project, but it's a challenge. One of these challenges, I think it was on Data that had posted it, but I saw it from Christina LinkedIn and she I think she's one of the judges. So that's how I saw it. And I just thought that it was such a cool subject matter. Disney movies, who doesn't love Disney movies? So that's why I try to make my own anime. I didn't do it in tableaux. I do it. I did it with Halaby and even the questions. So this one I was kind of, you know, ready. But still, I don't want to be too harsh about this subject. But one motivation for for me to do that, take part in that challenge was that I'm seeing these challenges and I see it on people just throw everything at this dashboard or reports like everything that they can find about the subject they will put in. And I'm I come from this whole other angle, like what was the question? Like, what did they want us to do? Like and only only report that. Yeah, I seem to be asked. Yeah, I tried to be as far as I know, I don't know if that's the right word.

Speaker1: [00:08:50] Yeah, no, no I like that, I like that approach like that. And by the way, it sounds like wherever you are it must be absolutely beautiful day because there's Data. It's nice, nice, nice harmonies in the background with those birds chirping. Put you on mute for the time being, even though I love the sound, those birds. But one thing I do notice a lot is any time A somebody is new to creating a project, they just using that as an opportunity to just dump everything that they know how to do into one notebook, which just kind of confuses the message and confuses the audience. Right. If you're doing a project and you've got your notebook up on GitHub, there's certain things that we don't need to see. Right. I don't need to see the dot info. I don't need to see the head of the Data frame. I don't need to see, you know, basic summary statistics. I just because it's something that you know how to do doesn't necessarily mean that it should be included in every single project, because at the end of the day, you want to have a message that you're communicating through your project. Imagine asking is probably the same thing when you're working on writing projects, right? Like a you don't throw every single word. You know, if you're trying into an essay, if you're trying to write an essay. Right. You have a point that you're trying to make.

Speaker2: [00:10:00] Yeah. One of the hardest lessons I had to learn about communicating a message. Right. And I think that's what you're getting at, whether it's through poetry or through essays or through fiction or whatever. Is that? You know, yeah, I had to learn that lesson hard, I mean, when I first went into grad school, like all my problems were just like the most, like, cliche, like overly written message, just to say something very simple. And I found that the best moments for the moments of really clarity that just like nailed down one thing very specifically. And I think that sort of relates here where like what is the task at hand? Like, what is it you're actually trying to communicate and what's the best sort of means of doing that? Not just like look at how I can communicate this in so many different ways. That was sort of I almost had, like, take my ego out of it was what I found was like once I kind of took myself out of it, I was like, maybe I do all that work to get to that distilled version of what it is. But that's what people need to see. They need to see the distilled packaged version of my thinking and my sort of wordplay and all that kind of stuff with poetry especially.

Speaker1: [00:10:58] And one one good way, at least four days on this project. I wonder I wonder if this is really the same case for for writing is you want to just have a plan of attack analysis plan set up front stating stipulating that these are things that I'm going to look at from this particular project. These are the methodologies I'm going to use. And this is my plan of attack. Right, that we have guardrails in place. And I'm sure it's probably different with creative writing. You I don't know, maybe maybe it's not. Maybe you do have like sometimes

Speaker2: [00:11:25] When you're stuck, it's good to give yourself some parameters, you know, whether it's I'm going to write this and sonnet form or I'm going to write this in five syllable lines or there's so many forms you can choose or restrictions or writing exercises. And I think like a lot of times when it's I'm getting stuck, that's what I turn to. And that's not being being similar in a certain way.

Speaker1: [00:11:46] So let's open the floor up. If there's any questions whatsoever on anything whatsoever, I promise I will try my best to answer it might not necessarily have an answer for you, but I'm happy to open it up. So what's up to Roger Peter Christoff, Cristoff? I know you're hanging out having a good good evening. I hope that's good beer that you're enjoying there. Make me a little bit thirsty man. Wait, wait until 4:00 pm here so I could, I could start to start, have a little little sip of some beer. Let's see if either Roger or Peter have any questions. By all means, go for it. The floor is open to you guys. I could try to stall and see if I can come up with the question that we could do a round robin topic on. Yeah, Cristoff go for awhile.

Speaker3: [00:12:28] And I noticed recently that I really enjoy the writing and I like to take advantage of Austin is now here and we had this topic I think like a month ago and you have suggested to watch the day you became a better writer I believe.

Speaker1: [00:12:46] Yeah. By Scott Adams. Yeah.

Speaker3: [00:12:49] But I believe Austin has more tapes. So my question is how to become a writer and how to be an effective writer, because you just mentioned that this is pretty difficult. And I know this is difficult to to really cut down to a single topic in a single article or anything that I'd like to post. And so how do you do that?

Speaker2: [00:13:19] Yeah, I think there's a few things. I think first is that whenever an area that you're trying to write in, whether it's fiction or poetry or essays or reports or academic writing, is to read people you admire in those areas and elsewhere to just reading it in broad strokes gives you a different sentence. Structure is different voices in your head. And then the second thing is like is empathy for your audience. And it's a that's a skill that you can develop. It's not just something you're born with or don't have, but it's really thinking about if my audience is is X, what's sort of the information that they need. And that can be especially writing a technical report or something like that. It's sort of like, OK, what's the next thing? What's the next thing and what's the next thing? Sort of working through it methodically. And then, you know, in terms of of writing, what I'd say is in practice and actually practicing writing is that if you're having trouble making things distilled down to their purest form or distinct meaning, what I like to do is I like to write everything out. I like to start by just getting it out, whatever it is, because if I'm if I start out focusing on I'm going to write the most concise sentence I can hear, then I'll just spin my wheels because I have a bunch of thoughts going through my head. So it's for me, it's an exercise of getting what I want out on the page and then just reading, reading a bunch of times, reading my own, writing out loud so I can hear read the language sounds overly written or stumbly and then just really going through and doing that over and over and over and practicing writing the same paragraph, a bunch of different ways to try to figure out what's the clearest distillation of the thing you're trying to say.

Speaker2: [00:14:53] And that's the practice, far of it. So there's a sort of the meditative part where you're reading people that that you respect. You're developing a deeper empathy for your audience. And then there's sort of the practice on the writing side where you are deep in the language and, you know, taking what? You've learned from the reading in the study and really just honing those sentences, having other people read them that you trust and yeah, it's a mixed practice of that sort of meditative and then actual on the page writing type of stuff. So I still struggle with getting my writing distilled down like my old boss. And I'm sure this will happen to us and write these products, release and release notes or announcements or blog posts. It's like you're burying the lead. You don't say the thing you want to say until the fourth paragraph. So one one specific tip, and I learned this from a poetry professor as well, is write a write a sentence that says that tries to capture everything you're trying to say in the piece of writing and then go from there. If you can capture the core thing in one sentence, then it allows you that can open things up a little bit and you've communicated the main thing and then you can sort of fill in the rest of the details there. So those are those are just a few scattered thoughts on that. It's been a long time since I've taught writing, but I'm happy to chat more about that, of course, also to any form of that kind of stuff.

Speaker1: [00:16:07] That's a really, really important set of tips and points there. And also we could take a lot for that as Data scientists because, you know, if you have a project up on GitHub, you should have an accompanying write up with it. Right? Maybe like a executive summary, I guess, for lack of a better word, but just something that that really clarifies for the audience when they get to your repository what it is that they're looking at and how it is going to be of interest to them. Thank you so much for those tips, Cristoff. I'll toss it back to you to see if you have any follow ups.

Speaker3: [00:16:39] And I do so because my topic is NLP because this is what I'm really interested in. And when you said that I should read a lot of stuff from people that I admire. So my question is, is this is this a good approach to people who write about NLP also? It's like the first step I should take.

Speaker1: [00:17:04] I would say it's a little different for that. Wouldn't you wouldn't you want to just practice because you could read books about NLP, but essentially it's going to be just the same thing. If you're looking just like a textbook or a textbook, it'll be the same thing. Just said different ways comes back to Austin on this one. But my viewpoint would be you probably better off just practicing, practicing, practicing and polishing away, polishing away, polishing as best as you can on that. But, Austin, I'd love to hear your take on this.

Speaker2: [00:17:35] Yeah. Are you interested in writing sort of a more like perspective on the the the study of natural language, or are you more interested in sort of like writing reporting on your own work in NLP? So I think it depends on the subgenre within that. Right. So if you're if you're if you if it's more of a high level, I want to think about how things are evolving in NLP. And I would suggest definitely reading folks who write about NLP sort of at that level, whether it's newsletters, the sort of longer form newsletters that people write that are sort of Sask. But if it's more of the technical stuff, I think you can you can think of reading, writing as more like models as opposed to a more philosophical exercise to like how do they how do they approach to step by step? How do they approach introducing the problem up front? How do they approach conclusions, things like that? Those are the two tracks that I'm seeing. And what you're saying is that I

Speaker1: [00:18:24] Like that and consider that. That's an excellent, excellent point, Chris.

Speaker3: [00:18:28] So I want to write about technical things, because I'm also quite at the beginning of the of of this journey. So I decided to write about what I learned. And it's I've I already understand a little bit of it, but my goal is to write to people who are just starting, who are total beginners. And it's like, I wish I read it a year ago when I started something like that. So it's it's for me a year ago.

Speaker2: [00:19:03] And yeah, I think in that case you can kind of free yourself up a little bit from from the expectations of like or like from being overly caught up on. Is the writing character does it in the right format. Exactly. If it's for yourself a year ago and it's for your own reflective purposes, I think the best approach is to kind of find someone you trust or a couple of people you trust to read that with you or to share if you can find other people who are writing and you can swap because I think a. drop something in the chat about editing other people's writing. And I think that's super important. I mean, one of the one of the core formats of any writing program is going to be the workshop, you know, sort of like passing each other's running back and forth, giving constructive feedback. And I think in small pockets, that can be really true for technical writing as well, especially if you're working with people who maybe are maybe on their journey or just a little bit further along and sort of have that perspective. They can help you. That's definitely, I think, a valuable approach, as well as forming a little reading groups and workshop type groups and just applying that to the more technical side of things. It doesn't AIs to be a creative writing workshop. Yeah, I like that.

Speaker1: [00:20:05] That the process of writing to yourself. From a year ago, I think that type of thing can really be beneficial to a lot of people in their journeys. I really like that approach. Um, so I just want to call it out. Your agent said he's trying to learn whatever he can. He has a few questions to chip in in the in the coming weeks and come through the questions now. And you might not remember them in a few weeks. And, you know, you might need the answer sooner than later. So, Raj, my friend. Go ahead, man. Whatever questions you got, man will happy to help in whatever capacity we can, because that's what we're here for, my friend. And you probably have to admit yourself because we're able to hear you. In the meantime, if you we'll give Roger the second or two to formulate his questions. And if I just feel uncomfortable, I'm meeting yourself. Go ahead. Just type it out into the chat with your question. Might be and I can talk about it here. Let's see what's up with either Peter or Joshua. OK, this is Peter.

Speaker2: [00:21:02] And I'm just wondering. So I'm an engineer and try to switch to Data science base. And there seems to be this thing about Stratifications helping to prove your level of knowledge in this area. And I know that this deejay doesn't always or other Kyle doesn't always be highly or stratifications, but I'm just wondering if it can actually be useful and what you would recommend.

Speaker1: [00:21:24] So it depends on what the argument you're making is or what you're trying to convince yourself of. Right. If you're coming in with the mindset of I have cert, therefore other people should view me as employable, I don't think there's necessarily a one to one core connection to that. Actually, I posted something on LinkedIn just like an hour or two ago where I said one of the biggest fallacies I've made through my career was the conflating the attainment of a degree with the acquisition of a skill. Right. Um, so certificate's themselves. Like, if your main objective is, look, I want to I want to go through a structured, formalized program. And at the end, I'd like to have a tangible benefit of that that I can put on my profile or whatever. Yeah. Go for it. But if you're approaching this exercise, the thought of, OK, I'm going to go through this program, get the certificate, and then now people will just know that I know what I'm talking about. They're just going to know what I'm doing and I'm just going to get a job.

Speaker1: [00:22:24] I don't think that is necessarily the right mentality to have because you have to have work product. You have to have tangible evidence of work, especially if you haven't acquired practical work experience. You need to do that in the form of projects. So I think, you know, a good friend of mine and if I could speak on his behalf and what what his point is, is that we and especially with Data assumed obviously a lot of people just assuming that if they get a certificate, then all of their woes will be alleviated and they will be able to get any job because they have this certification. That's not necessarily the case. So a certificate doesn't imply that you will you're necessarily even hirable. But if you have a well done, well constructed, a well thought out, interesting project where somebody can actually see evidence of real work, I think that's much more powerful and probably a better use of your time than just sitting through a series of lectures to obtain a certificate. I'll see if you have any follow up questions on that.

Speaker2: [00:23:29] Peter, I'm okay here, so thank you for that. I do understand that. I mean, I'm currently actually doing a program where I've done about four projects and I still have like two or three more to go. So I do understand that that helps to prove your skill level. I'm just saying that at times it's although it's not a measure of skill up in that certification, at times I think it might help. Or rather, I was thinking that it might help get your foot in the door or get you noticed in the first place to even be able to display that skill.

Speaker1: [00:23:59] Yeah. So, I mean, I think of it this way. If we have let's say I have 100 different versions of Peter Wright and I'm looking at 100 hundred different versions of Peter who are applying for this common Data science job. One job. Right. But the handful of versions of Peter, which have a certificate that will just signal to me that you've got interest in this field, that you're committed to the field. Right. So if anything, I'd say the certificate might be a obviously great way to learn. You should take that if your objective is to learn how to do something. But it, I don't think necessarily give you a leg up, because if I have one hundred versions of Peter and these hundred version of Peter, I'll have, you know, certificates. Right then the the versions of Peter that have a project are likely more more likely to get hired. So certificates are great to signal interest to signal me commitment and dedication to a field. But I don't think that necessarily means that you're automatically an employable data scientist. Cristoff, go for it.

Speaker3: [00:25:01] I'd like to add a little bit to what you just said, because I read this. This. About being in motion and taking the action and being in motion is like getting information about something. So imagine that you'd like to lose weight. So being in motion is like reading about healthy food and taking. And the action is like exercising and really going to a diet or eating healthy. And to me, certifications are more like being in motion or books are like being in motion or taking participation in such meetings like they want right now and taking and taking the action is doing projects for me. So taking the action is something that gives an outcome and being in motion. So and I believe that everybody talk about it, that projects are the most important. And I agree with that because without projects, everything that you learn and doesn't give any results, actual results. So that's my point.

Speaker1: [00:26:16] Yeah.

Speaker2: [00:26:16] Peter, thank you. Thank you. I do understand that. I think so. I come across a few opportunities where they actually ask if you have a background in computer science because you have a particular certificate or certifications. So in the analogy that you gave about 100 different versions of Peer that you pick the ones where projects, I'm just saying that at times I find that you don't even get to the stage of put in your project because you've been disqualified because the other versions had certificates.

Speaker1: [00:26:45] I mean, that's if you want to fall down that train of thought like you don't you can't control whether or not you progress through the process. Right. Like people can put whatever they want on the job descriptions and some organizations might value certificates over other organizations. Right. So that's one thing. There's the culture of the organization. Some might want you to have a certificate. Some might not even care. Right. But I'd also say this, that more than a certificate, what stronger is your resume? Right. Maybe maybe you got disqualified from the first round or disqualified from even being brought on for eight hour phone screen because your resume wasn't in good shape. Right. So that's an issue. So my philosophy of this is I know how hard it is in the job search because, I mean, when I was transitioning from statistics to data science, probably went on forty something interviews in a six month period. I mean, interviews, right. That doesn't count how many how many jobs that I've applied for. And what contributed greatly to my success was me fine tuning every part of the process, which I had control over. Right. So if I see a job posting, the job posting says they want this background or they want this particular certificate. Great. Doesn't matter to me. If that's what they want, I will still apply for if I feel like it's a place where I can make a contribution. Now, what I would do is make sure that my resume is in good shape, make sure that I've targeted people inside the company and sent them really well written reach out letters and letting them know of my skill set, things that I possess, skills that I possess despite their requirement for some type of qualification that makes sure that during the interview process I'm doing a great job responding to questions and and showcasing my knowledge and my skill set. Kevin, I see your hand up, so go for it.

Speaker2: [00:28:34] Yeah, it's just going to say hiring project managers, they have to have both, in my opinion. For example, I had people applied that got four years of college and their certificate. I wouldn't hire them. They had absolutely zero work experience. I look for people who had good work experience, plus the certificate to understand what a framework was and how to go ahead and do projects. So I always look for combinations of stuff.

Speaker1: [00:28:58] So, Peter, I'll turn it back to you if you have any follow up questions or statements on that.

Speaker2: [00:29:03] Thank you. Thank you.

Speaker1: [00:29:06] Yeah, by all means. Like, if you want a certificate, go for a man. Like, don't don't read what anybody else says about certificates like block you from going for it. But I think it would be a fallacy to think that just because you have attained the certificate that it indicates your employability in any way, shape or form. The employability is indicated through evidence of work, proof of work. Right. And proof of work is typically shown during projects or done during the interview process, where you have a take home challenge, where you've got coding challenges, discussing your previous, you know, work experiences and things like that. So let's see if anybody else has comments or anything at this point. Austin, anything to to contribute anything on this,

Speaker2: [00:29:55] To be totally transparent and the job search front and that sort of world. I don't I have stumbled back asteroids into all of my roles and sort of just by the connections and networking I Harp. To do through my brother and things and then sort of execute it once I got there and was, I think one of the I would say this, that one of the things that having taken advantage of opportunities is just being a curious person. I think people it's a really under under talked about skill set or talked about perspective and approach. The more curious, like just sort of, you know, jumping into my new role, a comment just like kind of being having my eyes everywhere, just understanding like what's going on with sales. Like I don't have anything to do with sales, but like what's going on there, what's going on in this sort of channel and this part of the business. And that curiosity has really paid off and allowed me to execute on opportunities that I think like good hiring. Managers are also looking for people who are curious and don't just have a set of skills that they're going to just execute all the time, but also a certain amount of curiosity and interest and, you know, ability to pick things up on the fly. I think that's super important. And those are the kind of people I want to work for as well that see those skill sets.

Speaker1: [00:31:00] And that's excellent point. And networking is super important. And maybe one thing you could do to get a leg up in the interview process is look for companies that you find interesting, connect with the Data scientists. They're hiring managers there. Don't immediately message them, asking them for a job, just like, hey, look, your company looks like it's doing awesome stuff. I'd love to learn more. Do you have, like, a blog, internal blog at your company that you guys have where you talk about cool stuff? Just like Comit has an amazing blog. Would they do so much knowledge sharing? Right. And maybe even try this for whatever next interview opportunity you have. Right. Let's say you see a job posted and that job looks awesome and you're like, man, this company looks awesome. I feel like I can contribute a lot of value in this role. But they're talking about they want me to have X, Y, Z qualifications. I might have qualification X and Y, but not the apply anyways. Apply anyways and then research the company, go on the website, learn about what they're doing, learn about the industry, learn about the competitors. Learn about how your particular skill set can help them solve the problems they are trying to solve at their company, in their industry.

Speaker1: [00:32:03] And once you have a good mental framework for what this company is about and the work that they're doing, reach out to a technical recruiter, send them a well constructed message, reach out to somebody who looks like they have clout and buy clout. I mean, you know, manager, director level and above of data scientist and send a message and say, look, man, I've applied for this job. I already submitted my my résumé through the online portal. But it's going to reach out and say, I found the work you guys are doing. Fascinating in particular. Here is this blog post that I read on your website that I found interesting. Or you could say in particular, um, here's an interesting problem that you guys are working on in your industry and how I can help solve it. You know, just anything to signal interest that goes along with it. There's the point I'm trying to make is there are so many other factors that are more important to your chances of landing your dream job. That certificate is close to the bottom of that list, I might add.

Speaker2: [00:33:01] One more thought to add on to that. And, you know, if it's a point of contention and point of some anxiety already or stress or whatever the word might be, I think if that is the thing that makes a difference, then it my job as it can be a job at certain points, but it might not be the one, you know, if it's the one for you, if that's sort of the core requirement that you're missing and you do all those other things and you expressed the interest and you show the curiosity, can you show the initiative? And they still and it's still like, no, you don't have the certificate then. Was that workplace going to be suitable towards your approach anyways? Maybe, maybe not. But it's sort of the thing like that I realized, well, like, you know, not every job is necessarily suitable for me and how I approach the world and how I approach it, I think.

Speaker1: [00:33:43] Yeah. Thank you very much. Ask him and Peter. But by any chance, are you a member of Data Sandström job? Um, just just curious, because if you are come to an officer, we can talk more in depth. Um, you know, I've got that I've started dating his dream job that you can attend. So if you are. Yeah. Yeah.

Speaker2: [00:33:58] Actually I paid for the A.I.s dream job.

Speaker1: [00:34:00] Remember all that. Ask me. I will make sure you're making good use of the office hours that we have in there and the wide range of expertize we got with all the mentors there. And we'll be happy to help you more personally during those office hours. So let's continue along here. I saw a question here from Roger that is saying he is at work. His main question is, as someone new in Data, science is amount of information or course or learnings available. OK, so you're pretty much saying that there's so much out there. I don't know what to choose, which path to go to to learn. All right, great. As one thing that came up was how to teach something which you knew years ago as not there. How can I find something like this as it will be really handy. So definitely follow a lot of really cool people on medium two of my really good friends that I highly recommend you check it out. There's Quyen Tran and Curtis Pikes there. They're amazing writers and they put out great content. So definitely check out check out their work. Um, but in terms of finding this. Which is the right pathway. I'll take a controversial stance here. It doesn't matter which boot camp you go to, doesn't matter. And look, and I'm saying this as principal mentor of Data science dream job. Right. We're not like an online education platform. We're not like a boot camp like that. We help you through the job search process and help you become a data scientist and level up.

Speaker1: [00:35:26] I'm saying that it doesn't matter which when you go to, because at the end of the day, we're teaching facts. Right? So everything that you're learning in Data science in terms of technical skill, those are all facts. Right. And they're very, very, very little room for interpretation of facts. Right. Maybe like the way somebody teaches it versus the way somebody else teaches it or whatever, but it's all the same. But I'm pretty much trying to say it is all the same. So I think optimize more for just finding the right type of people to turn to for questions, asking really good questions. Right. Which is ownership on you to help direct your course of study. Right. So when you talk about having to say the right pathway, well, first is to say very clearly what your objective is and what it is that you wish to do. Right. And that's the let's all say, because that's you know, I don't know much else about you, so I can't really give you more guidance or insight as to what the right pathway is for you. But before you start any journey, it's just get extremely clear on where it is that you're trying to end up and where it is that you're trying to to go can go for it. I don't know if I can just admit it to contribute or I'm just joined in this unmuted, but happy to have you. You can have never seen your name before. So how are you doing?

Speaker2: [00:36:45] I'm all right, thank you. I saw your LinkedIn post. So nice MOOCs. Let me jump in.

Speaker1: [00:36:50] All right. And will do. I'm super excited to have you here. I love the background man. Appreciate it. Thank you. Yeah. Man, how has it been. It's been good in real good man. So tell me a little bit about you, man. I do remember seeing a couple of messages, I think from you all LinkedIn, if I'm not mistaken, but I don't think we've actually got a chance to to interact much. So tell us a little bit about your family, where you at in your career, in your journey and in, you know.

Speaker2: [00:37:19] Yeah, no, I I'm currently in graduate school getting an MBA and the data analytics specializing in it and starting a new position as a political analyst soon. So that'd be pretty cool doing some data analysis as part of that position. So, you know, I'm just really looking forward to learning as much as I can about the industry and how to use the discipline. Right.

Speaker1: [00:37:38] And hey, so I used to be in the industry. I was a biostatistician for several years. I'm very familiar with the clinical trials and and clinical trials, statistics and designing experiments, stuff like that. So definitely fun. Interesting type of role. Um, so yeah. I mean at any point you got any questions about anything man. Just let me know and we'll, we'll, we'll get you some, some help. Yeah.

Speaker2: [00:38:05] Thank you.

Speaker1: [00:38:05] Yeah. Definitely. Also shout out to our aunties and his blog to check out and his blog. You can be a pirate dotcom. It's all about you. It's all about learning are um,

Speaker2: [00:38:17] Very, very beginning in the beginning stages. But yeah, it's been interesting to try it up.

Speaker1: [00:38:24] Yeah. I like the work you're doing there and I love getting home too, so that's awesome. So let's head to I hope I'm saying your name right here or Yanda or.

Speaker2: [00:38:36] Yes, I agree. Thank you so much. I hope you can hear me.

Speaker1: [00:38:40] Yeah. Loud and clear. Loud and clear.

Speaker2: [00:38:41] Yes. All right. Thank you. Well done. I've been following all the training on this DG. I just have a quick question. So I started working on the salary prediction portfolio project and then I realized, you know, the tests Data doesn't have target values. So how do we measure? So, yeah, I started with the training data and I could measure the MSI, got good values and all that. Now, how do I measure how effective my model is when I don't have target values for the training for the test Data. I, yeah. I just need some idea because I need, because I started that, I was excited. I got the idea. Well let's talk about values. How do I measure what I've done is good enough. Yeah.

Speaker1: [00:39:29] So just for full disclosure, for everybody wondering where all these Data job people are coming from, I shouted out to my Data says dream job community that do not open office OpenOffice.org just to get more people in here. So so that's that's why we got a couple of DST jobs in here. But to answer your question or know them. So just for some context for everybody else, as part of Data and his dream job, we've got a bunch of example projects that students can can work on to kind of develop their skills to see how a project should be carried out for that particular project. You're talking about the salary prediction project. There is actually a data set that is the called Why Test? So the X test, the X X test data set doesn't have. The target on it, because it's in a separate file called the white test, so there's there's that. And I mean, in general, you just take your fitted model, do not predict on the test, and then you'll end up with a vector of output values and you could use that to computer your MSA against your test. So hopefully that was helpful.

Speaker2: [00:40:28] Yeah, thank you. I just didn't find that whitest.

Speaker1: [00:40:32] Yeah, it should be in there. Go ahead. Post the message on our channel for that. OK. Yeah. But apart from the from the desktop project, any other questions I can help you with. Thank you for coming in.

Speaker2: [00:40:42] Yeah I'm good. Thanks. Thanks. I was surprised there wasn't going to be a notice out there. So when I saw this, I just didn't think

Speaker1: [00:40:49] So because I do I do office hours on behalf of my friends. That comment Emelle and these are open for freenode, the entire community outside of Data Sandstrom job as well. So, yeah, that that's why it wasn't announced in our slack. I just was trying to get more and more people to show up.

Speaker2: [00:41:05] All right. Thanks. Thanks.

Speaker1: [00:41:07] Right on. Joshel, my friend. I saw your unmetered there a couple of times. I definitely go and take the floor. Or keep it clean here a little bit, there's a little bit of a background noise, but I think we can manage.

Speaker4: [00:41:25] Ok, thank you, thank you. I would now. Yeah, well,

Speaker1: [00:41:29] I can I can make out I can make out what you're saying. So that's good.

Speaker4: [00:41:32] Oh, I think that's what the opportunity be there. So my question. Yeah. So I'm kind of starting out on this and stuff and I've worked around on some of the some of the tools sort of maybe Boatwright's interview and then some little of that. I never see. The part that I'm confused about is either to work on this or working on this level. And what my question is, at what point will this to be very much importance than the other? Or do they just level up on the. Yeah. This level of.

Speaker1: [00:42:14] Yeah. So let me just say I'm a picture unmetered in the background, but to rehash the question, make sure I'm understanding that you're saying that there's a bunch of tools out there for like data visualization and stuff. You mentioned Tablo, all tricks. And the question is more around, OK, which tool should I use and when is this tool going to be used? Is that the question? Joshua, let me know. Yeah, I would take the take that as. Yes. So, yeah, here's the thing, man. I remember at this job I had, um, they wanted me to make a dashboard, which I happily did, um, and I enjoyed making the tablet dashboard, but we ran across the issue was that for us to disseminate the Tableau dashboard, everybody needed to have some type of license for it, especially if we wanted to use it on our intranet trade. We didn't want to make it public where everybody could see it. And that was a barrier because we didn't have enough licenses and the licenses cost money. Right. So I guess my point that I'm trying to make is that tools like tricks and Tablo, their usefulness in the industry is definitely useful. There are companies that are using it, but not every single company is using these particular tools. So I would say that when it comes to dashboard and types of tools, I'm not like paid by Microsoft or anything like that. So this is not an advertisement for them. Um, you're probably better off using power by because that has a lot more a lot more market share, I think.

Speaker1: [00:43:46] Um, so I don't know if I'm answering your question, but if I was in your situation and you know, I'm really into data visualization, I would instead of focusing on a particular tool to use, I would focus on good practices for data visualization to make sure that the way that I'm presenting data in a visual format is effective. Rightside focus more on the effectiveness of, um, the the, the messaging, I guess, messaging, etc., but the effectiveness of your visualization rather than learning a tool, because at the end of the day, maybe you can learn Tablo or all tricks realistically within a week, maybe two weeks, and you can hit the ground running. But if you don't know how to create a good visualization, if you don't know the elements that make up a good visualization in order to effectively communicate the message, then, um, then you're going to be stuck anyways. And it doesn't matter what you're using because everything you create will not be effective. Right. So I would optimize more so for, um, for learning best practices, for data visualization and maybe sticking with open source type of tools. Um, definitely no. Keep doing stuff on Tableau public. I mean, I'm not sure what the market share is for Tableau versus Ultragaz versus power by, but PABI has got the added benefit that you can integrate Python into it. So, um, that's a huge plus. Joshua, I'll toss it back to you. If anybody else has comments on this topic, I'd love to to hear.

Speaker4: [00:45:09] So, uh, thanks so much. That was really helpful.

Speaker1: [00:45:15] And to answer your question, though, to answer your question.

Speaker4: [00:45:18] Yeah, he did answer my question. Yeah. That he said OK. Yeah, I guess so. Yeah. Yeah. Follow up question. OK, I haven't gotten into it. I would go to the stores and deeply so how do you how do they level up with. Oh we've been using up and that analysis. Celebrities such as SQL.

Speaker1: [00:45:41] Yeah. So are you Seabourne for any visualizations that I do. That's just my preferred thing. I don't know how to use Tableau. Um, not like the best idea, but I just don't use those often. I just use Seabourne for any data visualization that needs to get done. Any graphics that I make that I need to make ilu seabourne matplotlib just because it's open source, easy to use. And whatever my output is, I can see that as a image and incorporate it into a report or whatever, and we're good to go. Um so seaboards awesome. Daddy Seabourne. Um I think Dash is up and coming and becoming really popular as well, so maybe dash it something worth checking out as well. Um data visualization. Um it depends on what type. Roll that you're going for like me personally, I don't do much data visualization or dashboard whatsoever. I primarily primarily build models. That's most of what I do. So it depends on the type of roles that you're going for. But just having a good understanding of how to effectively communicate Data through a visual medium is probably more important than the actual tool that you are using. Seabourne is awesome. You Seabourne.

Speaker4: [00:46:51] Oh, thank you. Thank you.

Speaker1: [00:46:55] No problem. Yeah. Um, yeah. And again, I've heard great things about Dasch. Haven't got a chance to use it, but definitely check that out if, if, if you're in the market for uh, you know, looking for open source projects or open source visualization tools. Um, let's see if anybody else has questions. Go ahead and let me know right there in the chat. Uh, I see some comments here. Uh, Aerojet, thanks. Oh, you're welcome, my friend. Happy to help and talk to us about your experiences in power by and ah, I see some good comments here. First off, talking about the ad, you know, tooling, uh, go for it.

Speaker2: [00:47:30] Yeah, it's an interesting question. Um, like I said, if you want to learn something fast and I haven't used Tablo myself, but at least probably I was really intuitive to learn at least the basics again, probably if you really want to get to the master level, it's not that easy. But, um, I've been using it a lot at work for for over a year now. And the visualization that I did, um, on Friday, it didn't take that long with power because I know I know what I'm doing with power. Yeah. But as I wrote there, to understand the plots like what what what you're actually what's happening behind the scenes, it's been really interesting and useful to to learn a plot in. Ah, since I am the pilot now. So I have no idea what goes on in Python, but for our upload it has been really useful and learning that has been really useful to actually understand what what's going on because in power by you just, you know, you point and click and it's really easy to use, but there's no deeper understanding necessarily there. Yeah.

Speaker1: [00:48:51] Jeez you.

Speaker2: [00:48:52] So I use both. Yeah.

Speaker1: [00:48:53] Yeah. And I could, I can even say like for, for power by um like I've seen people at my work who are not Data analysts or Data people whatsoever that are higher level managers come up with some really cool looking dashboards on their own. So it is intuitive to use. I would say if if somebody who's not really Data analysts can make really effective dashboard with it. Um, but again, not an endorsement for any product or anything like that. My my big thing is and I think it'll always be this, it's focused more on principles rather than tools. If you focus on principles which are timeless, then you will always be an effective user of any tool. Right. This this is a course I took. I can't remember when I took it might have been twenty seventeen or something like that. It was through Coursera and it was a UC Davis course on data visualization. And the first half of that course was all about how to effectively communicate Data through a visual medium. And they had all these different principles like all these. They call them Gestalt principles, which is really I don't remember many of them now, but I remember that they did stick with me, um, and it was really quite, quite fascinating. Um, and just because Cam is a clinical trials here, I've got to pull up something real quick that I think will be interesting as it relates to visuals and clinical data. So give me one second, folks. I'll be right back here. I can't find the book, but it was right ahead right here. Or what on earth is this one? Medical illuminations.

Speaker1: [00:50:23] And this is all about how to use data visualization for using evidence of visualization and statistical thinking to improve health care. Um, so, yes, I was going for way more than the second. But, you know, I'll edit that out for the podcast. Everybody watching on YouTube gets to see me get up and walk. Yeah, this was a pretty interesting book that I had that talks about how to communicate Data through a visual medium. Um, so definitely check that out. And they go through some, like history of different types of visualizations, um, and different principles and stuff like that. So I highly recommend checking this out. Um, I think there's like a Data Data visualization podcast. I mean, there's obviously my good friend has some great content around that. She's kind of an expert in that area. Um, okay. Long enough. Yeah. To mention Kate. Yeah. Yeah. My my very good friend Kate who uh we're hosting a um a Kont Content Creators Award ceremony. You guys have not yet already registered for that or voted. For that, please do, it's the Data Community Content Creators Award asking if you can help spread the word with comment around this, I'll be awesome as well. This is hosting live. Yeah, it's like the combination of the People's Choice Award and the MTV Video Music Awards. It's it's going to be hosted live on LinkedIn and powered by you and your votes. I'm happy to announce that we have a very, very special guest house, John David, my friend, John David, host of the How to Get a Analytics Job podcast. What's going on, man? Have you seen

Speaker2: [00:51:58] Hey, how are you doing?

Speaker1: [00:51:59] Good, man. Good. Excited to see this first time we've actually ever interacted. And I'm happy that you one way

Speaker5: [00:52:05] Has we've been it's been a crazy week, so I don't know if you've noticed this, like really outdated looking wood panel, but I just moved into a new place

Speaker1: [00:52:12] To do the wood and set up wood panels or my shit man. I fucking love that stuff. Wood panels are awesome. I was trying to tell my wife that I wanted a wood panel my entire office, but she wouldn't let me really.

Speaker5: [00:52:25] Ok, so you don't think this looks outdated now?

Speaker1: [00:52:28] I love that stuff. OK, well,

Speaker5: [00:52:30] I've got like my monitor then my laptop and I've kind of like because we're starting a live stream now, which is a whole new whole new

Speaker2: [00:52:38] Experience.

Speaker1: [00:52:40] I don't feel excited. It looks good. I've been trying to it I've been trying to get my LinkedIn live certification for quite some time. I guess they don't think I'm a good enough content creator to go live, so I have to go recorded and share my stuff.

Speaker5: [00:52:53] I got a present for you. What's that apply for through your company page. Oh OK. I got it within a week.

Speaker1: [00:53:01] Ok, nice. Nice. I'm going to check that out. Somebody asking is blue yeti a default Mike. To what. What does that mean. Cristoff.

Speaker5: [00:53:09] This is a blue yeti right here.

Speaker1: [00:53:10] Yeah. Oh OK.

Speaker5: [00:53:12] Yeah this is pretty much standard. I think you've got a better mike. I mean you're your voice sound super crispy.

Speaker1: [00:53:18] Oh thank you. Thank you. Yeah. This is a 8r Tony. One hundred. It's actually pretty cheap. OK, so I'm like one hundred bucks but I really like, like this to set up. Um it makes it easy for me to be hands free. I used to have it where it was on a stand right in front of me and I always had to like talk facing down. It was not a good look. Um, yeah. So let's see if anybody else has questions. I, we just we were just hanging out today, John David, you know, typically how these things go. Same as my Friday office hours, happy hour or so. Other people just come through and if they have questions, we tackle the two questions, one at a time. But today people have just been talking about, you know, how to how to write, how to use data visualizations effectively. I guess that's an interesting question to ask you, John. So Josh was asking about the use of tools such as like Tableau and in the industry, if you're going to optimize for learning a particular tool, how do you decide which tool you're going to use? And, um, let's let's start there and then we'll go more

Speaker5: [00:54:23] So optimize in terms of like, are you talking about like a maximize your return on investment of like. Yeah, if I spend an hour on this it'll get me this much closer to the goal line. I'm getting a job. Yeah. I would say power behind Tableau or pretty much the lowest hanging fruit is you can learn those tools within. I mean you probably learn within a week and then all of a sudden you actually have a skill, and especially because a lot of executive teams, people are, you know, in their fifties and above. So you show them an interactive visualization and just blows their minds.

Speaker1: [00:54:52] Yeah, yeah, yeah. That same same type of sentiment that was having there was Pattabhi definitely because it's got such huge market share. And in terms of you mentioned showing off to executives Dashboard's, what other use cases have you seen for for these dashboards in the industry, aside from, you know, showing cool visualizations?

Speaker5: [00:55:16] So what do you mean by that? Like like different like like marketing. Sales.

Speaker2: [00:55:22] Yeah, graphic.

Speaker1: [00:55:23] Yeah. Let's talk about that. Yeah. Psychographics. Let's go there. Man, that sounds interesting.

Speaker5: [00:55:28] Oh man. This feels like a shameless plug for myself so

Speaker1: [00:55:31] I do go

Speaker5: [00:55:31] For it. We're so we just launched our own learning platform that's learned that sort of analytics dotcom. And what I've done is I've built out a prototype like retail sales case study course. And what I did was at the end of each chapter, I have a survey that asked you like three questions, like what's the overall quality? What was my presentation like? How well do I present the content? What was the content? Valuable and useful. And then an open source, just like, ah, open ended question of do you have any additional feedback? So psychographic data is essentially Data that you collect from survey. So it gets at thoughts, feelings, emotions. So what we're doing is we're using this course and the people, like the first wave of people, are creating the data source for the next case study course, which is going to be survey analysis like how to how to think through creating surveys, how to. Study them, and it's all based out of Tableau. So at the end of it, you have a specific business use case and then you have a portfolio that you can then use to kind of route the conversation during an interview process. So instead of saying I know Tableau, you can say, oh, yeah, I know how I know how to use Tableau to study sales data. In fact, let's pull up my Tableau public page and I can talk you through how I think about it.

Speaker1: [00:56:50] That's awesome. And this might be another use case that I had for using, for example, power behind dashboards. I mentioned at the top of the hour that actually before we started recording the show, that I'm working on this project at work that in it's a data management Data strategy data governance type of project, which I don't really know how to do because I'm a data scientist. Right. But part of this is involving us doing lots of interviews around certain topics to understand business drivers and benefits that people think they can get from a data management project. And we're communicating all that information visually through a power by dashboard. We have all the scores and we have there that we're using a radar graph to kind of show the multivariate relationships between how they prefer one benefit over and over, another side, so forth. So that's another use case for you, Joshua, like, you know, just different ways that you can use dashboards to communicate a message. Um, but, yeah, let's let's open up the floor, see if there's any other questions. I know it's a couple of people came in, a couple of people dropped off. Some people might have been shy, but actually Pew just came in. So if you have a question, go ahead of yourself and let us know what your question is. Adam, how's it going, man? Good to see you here again, Adam. If you got a question or anything that you want to chat about, go ahead and let us know. Otherwise, you could start to wind it down. But this is going to be this is going to be a cheap shot to ask a question. So I'm calling you man. Let me know. Let me know. First of all, just let me know how your week spent. How's it going? Pew. Pew, Shashi. Adam, how's it going, man?

Speaker2: [00:58:29] So, yeah, it's doing all right. And one of the questions that I have is I have and I'm sorry I asked this before, but how important is graduate education to becoming a data scientist? I've heard that you absolutely need it.

Speaker1: [00:58:49] Yes, a very, very interesting question. And this is a similar thing we were talking about earlier today. I don't know if Peter is still here, but Peter is asking whether certificates were important in landing a data science job and even had actually, you know, this is interesting that maybe this is me just putting this out there in the universe with with a LinkedIn post. But I had posted something on LinkedIn where I said that one of the biggest fallacies I've made was conflating the attainment of a degree to the acquisition of a skill and thereby letting my knowledge use me rather than me using my knowledge. Right. That being said, I do have a graduate degree in math and statistics about graduate training, but I can I can say with full confidence, I have learned far more outside of school than I have in school so I can. So I don't think in Data science graduate degrees are actually necessary. I don't think you need to have a graduate degree in order to land a job as a data scientist, because if you're if you're really serious about this career path, you're going to put yourself through self education and self learning and just continually upscale. Um, so what's more important is just your ability to learn new things and then effectively communicate that you've learned or understood something through through projects. Right. Um, so pause there and see if you have any comments or questions or John. If you want to chime in. Yeah.

Speaker5: [01:00:12] I've actually got thoughts on this. Yes. So my friend Moledet actually grew up with we just successfully pivoted her out. She was she got a degree in environmental science, but I guess a concentration in geology. And about two months ago, I convinced her that she should. Well, No. One, she's not happy. She's making like no money. So she's like, I am depressed because I've looked around and this is the only apartment that I can find. I was like, well, you know, my podcast is literally how to get an analytics job. So about two months ago, she started learning some basic level like analytic skills and she just landed a job two weeks ago, entry level Data science job where she does disaster relief. So I think what's interesting is that if you have an area outside that is not directly related to analytics, so the fact that she knows environmental science and has Data skills, that's a very unique brand skill stack. So think about like there's probably about a million people applying for analytics jobs, but how many people have geology, environmental and data science skills? So, I mean, that's a little bit of I don't know. Well, what's interesting is we're like. Actively developing like what? Like how we're thinking about things like as it's rolling out, but I think that it's there is almost like there's an overvaluation of hard skills. It's like, can you see? And then what's your perspective on things? Also, too, is valuable.

Speaker1: [01:01:40] Like, I like that you said talent stacking and skill stacking because I superimportant finally intersections of different areas of expertize that then you can become the best at that intersection. That is huge. But honestly, man, like here's the thing. Like any graduate program you go to, most of the syllabi are available freely online. You can look at a course and figure out what it is that you need to learn and put yourself through that if you really wanted to. But I don't think you need to have a graduate degree in anything to get a job as a data scientist. Now, that being said, if you are trying to become a research scientist or work primarily in research based roles, then I think graduate studies are probably more important, but only to the extent that they indicate that you could sit down for a longer period of time and study something and not get bored out of your brains by studying it. But I'll posit Adam and see if you have any further comments or questions based on what you heard here.

Speaker2: [01:02:34] Well, that's really good to hear, because my impression was that the Data science career path was that you would probably major in some STEM field and then you would do internships to get a Data analysts position and that. Well, really, you would probably not even in a STEM field in the domain that you wanted to do Data science, and then you would teach yourself the skills and aside as a minor or a double major, and then you would use that to get internships as a as a data analyst, and then you would do data analysis and then do your masters and upgrade to Data science internships. And then once you had finished your masters, you had years of experience as an analyst, then you might be able to apply for a Data science job.

Speaker1: [01:03:19] I mean, that's definitely one possible career path that people can take. But I've seen people land their science jobs with just the bachelor's degree and maybe a year or two of experience. If you are if you are an undergraduate program, definitely try as hard as he possibly can to get an internship just because you can learn what it's like to to work. But if you can't land an internship, then do projects and projects and projects and you don't even necessarily need to have a year or even necessarily need to have a STEM degree in order to get into Data science, do you need to know STEM concepts? Yeah, but I mean, just because you study a particular thing does not mean that you're incapable of learning something else. Right. Like one of my community members, Kourosh Kourosh Elizabeth, who has the philosophy Data project. He studied philosophy, PhD philosophy. And I think he did primarily ethics. And he decided to pick up Data science skills. And now he's got this really interesting project that blends philosophy and Data things just like philosophy Data project. You look it up and he does like this cool NLP thing with philosophical texts. Right. Um, that just goes to show that, you know, you can learn anything you want without having to go to having that particular form of training. Does that make sense, man? Like I guess what I'm trying to say is that I don't think that just because you majored in art history that you're incapable of learning math or stem related topics. Now, if art history major with no training in math or science or anything, try to get a job and analytics or try to get a job in Data science, that might be tough. There'll be a lot of self education that needs to happen for sure. I'm just rambling at this point. I'm going to pause and see if if there's any other

Speaker5: [01:05:04] Anime you may want to look into building on a portfolio.

Speaker1: [01:05:07] Yeah, yeah, absolutely.

Speaker5: [01:05:08] So having something tangible to say that this is the work that I've done and it's really fascinating once you get some distance between the work that you've done and then who you are, because I mean, that can get shared on social media and who knows who might see it. Like, for example, some of my students, Greensboro College last semester were on the golf team and they visualized their golf Data. And what's funny is that Kenji is a golfer. So when we had Kenji come on the podcast, they instantly connected. So, yeah, I think that think about it from like the human level as well. Like what is interesting about my personal brand that can connect with other

Speaker1: [01:05:44] People and and he has a good anecdote here, man, and he talk to us real quick.

Speaker2: [01:05:48] Yeah, I went to university when I was young, young, like most people dropped out because of life things. Now I've been working. I'm now working as a business development manager and trying to go towards the Data Data career. But I'm already using everything that I'm now learning in my current position. And it's just I haven't learned math since since, um, senior high, but. It's just tougher, but it's not impossible to to learn things that I need now just means more work.

Speaker1: [01:06:34] Yeah. So, I mean, I know we give a lot of good general advice here by Adam. Like, what's your current situation like? Are you in grad school or are you in school? Like what's going on with you.

Speaker2: [01:06:41] I am in school doing physics. I taught myself PyCon by Data stacks. I get learn the equivalent English accent machine learning libraries built Data pipelines for code for America or I should say architected and did the research for them. Now I'm getting the core math of another code for America one. And I'm starting to teach myself neural networks and oh my gosh, I have a whole bunch of Sims that I've written and I'm trying to start this and I'll consultancy with my friend. And yeah, frankly, I just feel super unqualified

Speaker1: [01:07:16] And I think I think you're definitely qualified. So you're not getting your qualifications despite having a piece of paper. You're getting qualifications by doing things which is far more important, your skin in the game at this point. And so I think you might just be selling yourself a little bit short there. What do you think, John?

Speaker5: [01:07:34] Yeah, I was going to say, Molly, my friend, that we helped get an entry level Data science job. She doesn't even know how to code in Python like she

Speaker2: [01:07:43] To do the Data science then are

Speaker5: [01:07:46] They are training her on the job. It's interesting because I feel like you're OK. It's an interesting frame because I think you have some anxiety around getting your first job. So your idea is to like acquire all the skills. But what's also an interesting take is that Molly is kind of like tabula rasa. She's a blank slate so they can train her on how whatever she wants, on whatever they want her to learn. So she I mean, it's funny because she went through this imposter syndrome and you can see it on the podcast where she was like, I am totally unqualified for any of these jobs. But yet they had she had multiple recruiters actually reach out to her once she started getting more active. Like, I guess my advice to you is maybe spend less time acquiring skills and more time going out and making relationships, building your personal brand, like get out like like how to kind of articulate that. So, like, you already have the skills, like you're pretty much good on that. Like you have more technical technical acumen than I do now.

Speaker2: [01:08:46] I can't like I feel like if I'm going to do Data science, I should be able to be placing at least in the upper half on Kaggle, pretty regular.

Speaker1: [01:08:54] Kaggle is at the real world, my friend. That's not a real Data science is like that.

Speaker2: [01:08:58] I know that. Of course, there's a lot more exploratory data analysis, consulting domain experts talking to business people, presenting your data and working with other people. But if you can't go out there, sit down with Titanic, if you can't do that, it's a pretty good litmus test for whether you can do it or not. And I got a

Speaker1: [01:09:18] I don't know if I agree with them. And like I mean, I probably would not do well in most Kaggle things, but I'm a pretty damn good data scientist. I mean, I make good money from my organization. Right. I'm recouping a lot of topline revenue with the stuff that I'm doing. And that's what matters, right?

Speaker2: [01:09:35] Like I faced empty Data values and straight up chote, I couldn't get extra boost running. I couldn't get freaking random forest classifiers going. And when I see papers out there talking about what they did, they're talking about ninety six percent accuracy. Ninety four percent record. Like, yes, I should be able to do that. It's like not being able to be easy as a programmer.

Speaker1: [01:10:00] Yeah. So not all data sets are going to be able to support that type of predictive power. If you end up in a situation where you are fortunate enough to have data that is that can support that type of predictive power, you can get those results. But more often than not, in the real world, you won't have data that can support that type of predictive power. And you have to just be happy with doing better than chance. Right. And maybe you probably have to look at something called Koans Cappa for a classification. I just seeing am I doing better than chance and really going for a pre contrived accuracy. A recall score probably is not the right way to approach your problem. I would say you should look for establishing the simplest possible model baseline model that will get you a good result and then from there incrementally trying to get better and better until you notice that it's kind of flattening. Right. And call it a day, deploy it, make money for your company and move on to the next project. Yeah, I think

Speaker5: [01:10:53] That that make money. Yeah, I think that. Adam, how well do you know like business. What's your business acumen like. Like, like how, how well can you apply analytics, data science visualization, machine learning to either increase revenue or cut down costs or reduce risk.

Speaker2: [01:11:09] Heck, if I know I've never done it before.

Speaker5: [01:11:11] So that's where I think you should spend some more time, like kind of divert some of that energy of like because it seems like you're going super deep because but also, too, I realize I'm just like a Data. I'm like more of like a management consultant with a little bit of Data stuff. So I know that I probably should even be on the machine learning office hours,

Speaker1: [01:11:28] I that we're happy to have you here, but I'd say, did you start applying for jobs, just start applying for jobs and kind of get out of your own way and out of your own head and get a resume, put it together, apply for jobs, go through the process and see what happens, like find an entry level role, apply for it and get real world experience, I mean, and then simultaneously skin in the game. And Nassim Taleb, to understand that intellectualizing things is far different than actually doing. Right, because I'm going to learn more by doing on the job than you are going to do by doing a research paper or doing a catalog competition.

Speaker5: [01:12:05] So I'm so glad that you like Nassim Taleb do. Oh, yeah. Yeah. I literally reference skin in the game of like you have to get out there and has to count. Yeah.

Speaker1: [01:12:15] Yeah, absolutely man. So in the same time that guy is technical in certain series which he made completely free on archive X. I'll give you guys a link to it. Um, statistical consequences of fat tails right here is a link to it. It is hilarious and typical Nassim Taleb fashion, but it's a technical book on statistics and the misapplication of statistics. He also has a really, really interesting open book that you can find, um, on on I'm sorry. I just need to have your hand up. Go for it.

Speaker2: [01:12:53] Or is Adam. I was just curious how often or in what ways do you reflect on and this isn't always true, but I mean, are you writing about what you're learning at all? Are you documenting are you because it sounds like there's just a lot going on, a lot of different things piling up. So how are you kind of like keeping track or like reflecting or like investigating what feels like a failure to you, I guess? Are there ways that your approach to doing that or is this sort of like just trying to acquire as much as you can right now? I'm definitely not really keeping score at all and just trying to get as get my hands on as much as I can and get to the point where, like, I could sit down with a dirty Data set and without any mess or fuss. I mean, obviously any unnecessary mess or foskett it's going to be messy and fussy, but without any unnecessary fuss. I get that Data said to, you know, state of the art performance, considering what I can, you know, what I can do with that. If it doesn't have much predictive power, there's not much doing. But I should be able to you know, if you gave me something similar to the diabetes Data, I should be able to get pretty good predictive power if it's crummy, really crummy Data that doesn't say much. I should at least be able to squeeze out 80 to 90 percent of the goodness over a couple of weeks. Right. So if I can't do that, then I'm not really doing data science and people don't really need someone who can just like get worse than whatever else their team can already do. Right. I mean, you don't have to go out there and beat soda, but you should be able to at least implement it reasonably quickly.

Speaker1: [01:14:34] Yeah, I don't know if I completely agree with everything you're saying there. There's the I mean, building models, predictive models is one piece of data science. It's not the entirety of it. There's a lot more that that you do on a day to day basis. If you find yourself in a position where all you do is build models, then that's awesome. That's probably more that would probably be more of a research type role, right? Yeah. So so do just apply for jobs. I challenge you to apply for like ten jobs this week, entry level jobs and just see what happens, go through the interview process and whatever this know, this notion you have of I should be able to do this, I should be able to do that, just get rid of all that. Just get rid of all of that, because there's no there's no there's should that it's just can I can I show up and solve a problem that is a problem worth solving that is going to help this company do one of three things, like John was saying, reduce risk, reduce costs, increase revenue. And if if those three things don't sound like something that you are interested in doing, then by all means, go to graduate school, get a PhD and work in a research environment where you are completely disconnected and removed from the real world and not really doing anything business related and then just kind of intellectualizing things because then you.

Speaker2: [01:15:51] Yeah, I would just say in the moment it feels like there is like I don't know the specifics in terms of how you're feeling about this, but I'm just thinking of a lot of what I've been through with, like, you know, constantly getting rejections from my writing. It's like a lot of it is about contextualizing what feels like and what is failure and figuring out how that contributes to what you're learning and how it's building on itself. And I think, like, that's why I always I mean, I always turn to writing to do this right. Or if I'm experimenting with something and documenting those experiences and trying to. Stand and sort of debrief on why they went wrong or why they didn't go the way I thought and sort of create a plan of attack to go to the next draft of the column or the next round of the experiment and just finding ways to contextualize what feels like failure. So it doesn't feel like you're just racking up bills, but you're actually doing is sort of like the other way because like there's no failures only like these learning experiences or sort of cool around that that I can't control right now. That's pretty pretty well known, I think. Justifiably so. So I think like figuring out ways to contextualize what feels like failure in a bigger picture of.

Speaker5: [01:16:57] Yeah, I think the quote is I think the quote is there is no losing. There's either winning or learning something

Speaker2: [01:17:04] Like that along those lines. Exactly. Exactly. And I think just like so that's just always been super helpful to me as it's not if you feel like you're racking up bills, it's like there's just there's something to be gleaned from that. And that's how you how do you extract that and then learn how to whether it's writing about it or whether it's having these conversations or whatever it is, it's never quite as simple as like it's a success or a failure. There's always something in between that and figuring out how to index that, catalog that and categorize that. That's different for everyone. But I think that's a super important skill that keeps you from, like burning up and getting overly frustrated and overly down on yourself. And I say that because I know how that feels, especially someone like non-technical who constantly is like working with people who have this like very highly technical skill skill set, or I feel like they're way beyond me in certain ways. And it's it's that's that's been a long game for me. But I think that that's a really important skill to develop over time. And, um, have

Speaker5: [01:17:58] You ever heard of this concept called a vanity metric?

Speaker2: [01:18:01] Yeah. I mean, it's basically things that people show off in order to impress other people, but which don't really matter. Like recently when machine learning, science papers, Data research Data science papers have been published. They've been posting these little incremental accuracy improvements of the speed sort of Harpreet Sahota, but no one can reproduce it. And so really now if you have something that beats the current benchmarks, I think it's more of a vanity metric at this point. The improvements haven't been huge lately, even ignoring whether they can be reproduced. And it's really just an entire ecosystem that emerged around that being the litmus test of a good paper rather than anything else, because it was easy to check. And now the machine learning science community is in crisis. So other things might be, for example, if you don't have a lot of context, you might be talking about, I don't know, ARPU, annualized revenue.

Speaker5: [01:18:55] I think I just wanted to use the vanity metric is like a mental model or mental construct. Essentially, I think that you you acquiring your KPI, like your key performance indicator is what coding language or what projects have I done? And that feels like it's self soothing to you. But really, I think the metric that you should be using to track success or wins for a week is how many interviews that I go in. How many analytics managers have I met? Like how many like going out and getting out because it seems like it's all like in your head right now. And I mean, it's it's impressive. I mean, you're more tactically sound than I am, I think. And I mean, I've but I've been in this space, although I guess like, I'm not a data scientist, so I feel like I'm but I have used Data to work on projects that have been worth like five million dollars for one of my clients. So I think that and I also manage data scientists. So I guess it's I just have a different perspective or view on it. I mean, what do you guys think now?

Speaker1: [01:19:56] I like that. I like that. Make me very metric number of books, read value metrics. How many people download my podcast. Things ultimately don't have an indication that I can't act on it. I can't do anything to to improve it. It doesn't really tell me much. I agree with you. I agree with everything you just said, John David there. But I mean, let me flip this back around to analyze what is it that you are trying to accomplish and achieve or do you just want to get doing to get a job and and start making.

Speaker2: [01:20:21] Oh, God, turn that sound off would be better. Yeah, well, I think I want to get to a place where I don't feel like I would be sending out a hundred resumes to get a callback.

Speaker1: [01:20:37] That's everyone. Like that's that.

Speaker5: [01:20:39] Hey, Adam, can I tell you about can we talk about Molly? Molly has no she has worked in the job for a year where she basically copy and paste Data points from this file to that CSV file. She went on to interview is the second interview. She got hired for an entry level Data sign shop. So, I mean, it's maybe that's an anomaly. So I don't know. But it sounds like just get out there. I mean, I don't think the interviewers are saying are looking at the these resumes and saying what what is this person doing applying? It's I think that just

Speaker1: [01:21:13] Go just

Speaker2: [01:21:15] Get them. They haven't. Yes. And it just automatically if you look at their job requirements. Right. So you type in Data scientist into like. Yeah, right, first thing that comes up positive.

Speaker1: [01:21:27] Let me let me let me tell you the proper strategy to do this right. So if you just apply for a job and click send and then do anything after that, you're just giving it up to chance. Right. First of all, you need to start with the assumption that any job that I apply for, there's probably only a one or two percent chance. There's only one or two percent of future realities where I end up with this job, have that as a base rate for any job that you apply for, because everything is a numbers game. But when you are in that job search process, there's things that you control that you should optimize for and things that you cannot control, things that you can control. Your resume sending messages to technical recruiters inside the company, sending messages to hiring managers that have director level manager level type of titles. Right. Sending them good messages consistently following up maybe at a maximum of three times and send them a message once every five days from maximum of three times. And then if you don't hear anything back, structuralist move on to the next one, but you will have to apply for many, many jobs. But if you just if you just hit submit and hope and pray that the gods are going to pass your resume through their filter, then you're just not controlling what is in your control. You're just giving it up and hoping that something positive will come out. So take steps that are in your control to land that interview.

Speaker2: [01:22:50] So I have a question about that messaging, that thing. If I were in their position, boy, would I hate that and think the people that did it and

Speaker1: [01:22:59] I'm not a recruiter, though, man. Right. Like, what is a recruiters job or recruiters job is to get people into roles that they have open. So you're not them, right? You're not them. Yeah. So message a technical recruiter whose job it is to fill roles. That's their job, man. So getting messages right

Speaker2: [01:23:19] Was a dream come true for a technical recruiter. You know, it's like I did a little bit of that work for the dole. If I could have gotten someone who, like, understood the company we were hiring for, like a job role, the company presented something coherent. And it wasn't me going out and doing cold calls. That would have been my dream day if someone came in on my arms and was like, hey, I want I'm interested in what you all do and was specific and kind and thoughtful like that would have been a dream. That means like, hey, maybe I have a pathway to some commission here. Like if you think about it from their perspective like that, it's a sales job, like they're trying to sell, like they get commission on that. It's not just like that's money walking out the door and throwing itself into the dream. Yeah. So I don't sell yourself short on that. And, you know, maybe they and you can't you can't assume what kind of edict they have or what sort of how they approach their job. You know, different recruiting offices work differently or some as an agency versus in-house as is like.

Speaker2: [01:24:11] But if you I think if you approach folks genuinely and you actually target with the way Harp these folks are talking about, it's like, you know, there's no you don't get put on some black list, that's for sure. You're just looking towards the and maybe the time isn't right. It is or it isn't. But I know I would have loved that as much as a technical career. So just a little follow. So someone not long ago and she said, OK, so you're about to graduate. I'm like, now I'm going to still be weirdly stuck in school for a year. You know, I took all the classes and she said, oh, well, they don't have a degree. Maybe I can say that you have all this experience. I said, no, this would be my first job. And she said, well, I don't know if I can help you that so. And this is one for one of the biggest agency recruiters out there. So no no diploma, no experience, no way.

Speaker1: [01:24:57] Yeah. So just like I said, numbers game, keep reaching out to people. And sometimes the technical recruiters are like, you know, there are technical recruiters at Facebook, at Amazon, at Company X, Y, Z. They have their internal technical recruiters. So whatever the company is that you are trying to apply for, AIs, LinkedIn go to the company page, just type in technical recruiter and find the internal people and message them. We're going to start wrapping up here. So I'm LinkedIn

Speaker2: [01:25:20] For all of his attention and

Speaker1: [01:25:21] Help. Yeah, no problem. Send you a link right now to to vent. You know who the investor is. He's a good friend of mine. He's got this channel on YouTube called the High EROI Data Scientist, and he's a giant in the field, an absolute legend in the field, somebody I truly respect and admire. And I think you will do well by checking out some of his content. Um, the thing. Well, you guys. So, yeah, awesome. Seeing conversations will be valuable regardless of the outcome of the recruiters. Absolutely. Um, well, thank you guys for hanging out. I know we went quite a bit over time, probably like one of the longest OpenOffice.org hours we had. So I appreciate having you guys here and sticking with it. John David, my friend, thank you so much for hanging out with us today. I appreciate having you here taking time. I just get to hang out with this. Um, Austin, it's always a pleasure, my friend. Thanks, guys. Make sure you tune into the podcast or release an episode, uh, just a couple of days ago with the author of Liminal Thinking. His name is Dave Gray. He's written a number of awesome books. So definitely check those. Out and check out the the interview I did with him next week about an interview released with a good friend of mine. So it's more of a personal interview with just me and my friend talking. So if you want to learn a little bit more about me than that, be a good opportunity for you to do that. Guys, take care. Have a good rest of the weekend, remember? You've got one life on this planet. Why not try to do some big cheers, everyone.