 Oh, everybody. There probably isn't anybody here yet. But I'm Monica Wahee, and I'm starting the stream a little early, just so I can give people a chance to connect. And so I can make sure that this is working right. Because restream, the software I'm using, is upgrading itself, which is good, but it means that I'm on a learning curve with using it. So hopefully I'm getting better. I see there's one person on the stream, welcome. So hi, this is our topic for today is I'm going to give you advice on researching companies in data science. What I'm doing now is this thing I would always do when I teach, where I'm waiting for everybody to come into the room if I'm teaching in person. But now I'm online, so I'm waiting for everybody to come in. I just start saying stuff that you can miss, if you miss it. Because then it's like I just repeating stuff other people know. So it's not like they're missing anything. But people can hear me talking, and I'm saying something important. So our topic for today is about advice on researching companies in data science that you think you might want to work at. So although I was originally aiming this at job seekers who are new to data science, I realized that I'm in data science. I have my own business now. But before then, I was a job seeker. I was always changing jobs. If you get to know me, like I remember one of my colleagues, like when I was 40, I was talking to her. And I told I was talking about different jobs I had. She says, how old are you, 60? Because I had had so many jobs. And it was so funny when she said it. So when you're actually building a career in data science and you've had a few jobs, it's actually really It's almost more important to research the companies you're thinking of working at before you think of changing jobs, because you're going to get rid of your old job. And so maybe it sucks, but I've seen things suck pretty bad data science. So you want to be able to do enough homework. See, sometimes you can get a good job. It's not a bad job, but it's just not a good fit for you. For you, like everybody around you, like I remember talking to one of my learners. He was in business intelligence and he had a job. And I said, well, what's wrong? And he's like, I like all my coworkers. I like what we do, whatever. I just don't, I'm just not inspired by our work, you know. And, you know, I don't know what to tell him. Like, would you like to do, to work on a different topic? And he's like, yeah, I'm inspired by this kind of data, but not that kind of data. I think he was an insurance. And so I was like, okay, well, then you have to change jobs, you know. And so it's like, if you're like, oh, I don't care what job I have. And then you get a job in the insurance. And you're like, this is really boring. Insurance data is really boring. Then I'm like, okay, this is exactly why you should be on this live stream. Because you should know you're gonna think an insurance job is boring before you apply and go through all the BS of getting the job. So let's see here. Some more people are joining. I see three people. Hello. If anybody wants to start chatting, that would be great because I can see if the chat works. So, but anyway, welcome. I'm just kind of BSing a little bit. Spaghetti dude, you're back. Thank you. And LinkedIn user is here and says, oh, I'm gonna put this one up. Hello, thanks for hosting this conference and sharing your advice. You're welcome. And feel free to ask questions. The main reason why I have a live stream is like I don't know exactly what your questions are. So, oh, it's lungs detoxification. My best fan. I love lungs detoxification. There you are and spaghetti dude. Okay, so let me take spaghetti dude says, can I take your opinion about something that is not really too related to data science? Well, I'm so curious. Of course, you gotta ask now. Okay, so you ask and I'll put it up there. I know how to put your questions up now. So, spaghetti dude, you type out your question and I'll put it up and answer it, okay? And then lungs detoxification says, I'm gonna put this up here for you. See, I'm getting good at the software. Ma'am, my question is I am 30 and I am new to this field. Past six months, I learned Python, Panda, Skill, Map.lib, what should I do next? Okay, and like I need a track. And then he says, he could be, she, big fan, man. Okay, I will tell you what is your problem lungs detoxification. It's that you don't have a topic, right? And like you don't have a field that you're in. So I'm gonna ask you a question. This is gonna be like a live interactive session. I want you to throw out a topic lungs detoxification in the chat that you know a lot about. Either it's a hobby you do, or like your parent does it for their job and you do it like if you were somebody whose parents own a restaurant and you grew up working in a restaurant. You just throw out a topic. You don't have to explain why you know it. Just throw out a topic you know a lot about. Like sports fans, maybe you know a lot about sports. I know I was a fashion designer. I just know tons, tons about fashion and about sewing and stuff. So I need you to give me that topic because then I will answer your question, all right? And then, so I'm gonna wait for lungs to answer and then I'm gonna go after what Spaghetti Dude is saying. Okay, Spaghetti Dude. Spaghetti Dude's question is related, is somehow related to the long question. But I studied computer science and then I served in military for years and left it. Oh, you know what's interesting? I worked for the military for a little while as a civilian, so I kind of know about that. And I was thinking of going back, here's what I would do. I would be a data scientist of military data, okay? Are you familiar with the MDS? The military data, God, I can't remember what it stands for. It's actually, lungs, I'm gonna get back to you. I'm gonna show you, I'm gonna go look up this webpage and then I'm gonna show you guys, I've learned how to do this, okay? Because I wanna talk about this MDS, let's see here, data set, the MDS data set, okay. Or no, this is military data centered. I'm trying to find this place where, I guess they moved, here it is, military, the military health system data repository, okay. Now I'm back here and I'm gonna share my screen here. And I think I can do this, let's see here. Okay. So here it is, here's my screen, right? So this is the military health system data repository and I'll even put the link in the chat here. I don't think you can use that data, anybody out there. What you have to do is you have to be with a group of people that can request the data, okay. So spaghetti dude, let's say, you obviously have ties to the military because I know what it's like, if you're in the military, you still have friends and every veteran's hanging out and stuff. What you would do is you would say, hey, I wanna work on MDS data, that's what this, or MHS data, like they'll say call it MHS, MDS, military data system, MDR, I've heard it called MDR. Just say I wanna work on this data and then let me see if I can show you. This is actually a federated repository. Oh, I didn't really officially open the live stream. So let me do that just really quickly here. How do I do that here? Hi, everybody, I just wanted to be big. If you've just joined, well, actually, there's fewer people here now than there were before, but if you just joined, this is, excuse me, Monica Wahey and today's topic, we have data science chat, today's topic is advice on researching companies in data science, but you can ask about any question. And right now I was answering a question for spaghetti dude who had been in the military, is studying computer science now and kinda wants to get into data science. And so I was just, because I've worked for the US military, I was showing spaghetti dude, who I assume was Guy, given the name dude, that there's this place, which actually has federated data sets. Let me go and look if we can see where all the data are. There's a data dictionary in here that's a big, you know, let me, oh, like here, Tri-Care Encounter data, which is just one example of the data they have. And so if you're familiar with clinical encounters that's in my field, oh, Mohammed, I'm so glad you're here. Thank you for joining. This is one of my friends on LinkedIn. Thank you for coming. All right, so what I was talking about is this military data that's available, but it's not widely available to everybody. You kinda have to be involved with people connected with the military. So spaghetti dude, let's say you became a freelancer like me or you joined some sort of consulting firm that works with the military, which probably wouldn't be hard for you because you're probably a friend in that field. And, but I, so spaghetti dude says, but should I be knowing health stuff? And that's a good point because I'm pointing out this military data repository. There were two main repositories I used when I worked at the data warehouse. It was this one and DMDC, which is the Data Manpower Defense Center, which had data just about the military, like deployment and like accession and those data sets. So this is what I'm answering you. You know the business rules of the military. Like you know all those data sets that the DMDC has. If you are working on a team that's using these health data sets also, you kinda know something because if I were to tell you, if I were to say, what's it like to be on TriCare? You would have an answer. You'd know what it's like to be on TriCare. If I were to ask you, what's it like to go to an MTF? You'd have an answer. You'd know what it's like to go to an MTF. Already you know way the heck more than anybody. Like I was on these teams of civilians who were like, what's an MTF? I had to go to Fort Bragg and visit an MTF. MTF is military treatment facility. It's an actual like facility that you can't go to unless you're in the military and you're probably living on the base or you're married to somebody living on the basis they're in their family. And so I'd never even heard of it, you know? And that's how like dumb I was. And I was trying to run a data warehouse in the military and not knowing that. So that's what I'm kind of pointing out, SpaghettiDude, is that's how smart you are. And that's how important subject matter expertise is. So SpaghettiDude, you could probably go to any of your old friends who have consulting firms or whatever, especially veteran owned, that are working with the military in doing, even if they're not doing data analytic contracts, if you are affiliated with them and they have a big contract, you could say, oh, can I do a little data analytics for you? And you can easily get in there. See, I can't, because I'm a civilian and being a civilian really hampers you in trying to do that. But being a veteran really escalates your chances that works. So I've tried to partner with like veteran owned places that do that stuff to try and get in, but you'd just be one of those people. So that's what I would recommend to you, SpaghettiDude. Why don't you try that and let me know what you find. Like let me know if you can find, because there's so many of these little veteran owned firms that are totally contracted with the military. And I think it's a good thing, because obviously they understand the military. So that's my advice to you. And then I wanna go back to lung detoxification because lung detoxification said wine brewing. So this was the problem is lung detoxification wants to apply learning from Python and other courses so that, or actually I'm telling them that they should apply that to analytics. So wine brewing is something that lung detoxification is into. So what does lung detoxification know about wine brewing? Well, I already, I know a little bit about wine, but I don't know much about brewing it. But I would say that lung detoxification knows about different grapes and about how to brew wine. Like what are the steps? Probably knows a lot of terminology in there. Now what I'm gonna do is I'm gonna go look for a data set that you might analyze if you knew about wine brewing, right? Like I'm just gonna go out here and go, okay. Download data about wine brewing, okay. This is like a total stunt, okay. Ha ha ha, look at that. Day 22 wine data sets. Okay, I don't know what's in these wine data sets. I don't even know if you can download any of these. I've heard of data world. Let me go back to the chat here. Does anybody know about data world? Like as anybody, you know, but I, this is kind of what I do as I start by going on, like those of you who've read my casino portfolio project example, that's literally what I did. I was like download casino data, Massachusetts. So Mohammed says, but getting data, let's see here. So Mohammed says, but getting data from veterans or their colleagues has much restrictions, right? So isn't it difficult even through contract with veterans? Actually that's what I'm really good at is negotiating data use agreements. So what happens is within the military, they transfer data back and forth with no problem, right? If you become a contractor for the military, they set up privacy between you and them, like theoretically, or you know, like with safeguards. And what I mean by safeguards, they make policies about your relationship. They also say things like you're not allowed to download this or that, you know, like what I would do at my data warehouse is I'd figure out what data everybody wanted and then I'd, okay, let me figure here. I'd figure out what data people wanted. I do data curation and I'd sit down with them and I'd figure out what they did they wanted from my repository and then I'd just give them the ones they needed to be identified and stuff. So the sort of big picture answer to your question is it's hard to set up the relationship but once it's there, it's easy to share data. So what I'm telling lungs detoxification to do is not, don't worry about setting up the relationship, just go be friends with a group that has already set up that relationship or you can go into the military but she's already been in the military. So that's not it, so let's see here. So yeah, so lungs detoxification, no. Yeah, I worked on wine data available on Kagle. Okay, so that's a good point. So I'm gonna go, let's see here. I don't know how to do this here, maybe that's why. I'm kind of, maybe I should stop sharing. Let me stop sharing my screen. All right, so lungs detoxification says I worked on wine data available on Kagle. So now I'm gonna ask you a question that might be hard to answer, okay? But I want you to answer it. What was the research question you posed to the wine data that you found on Kagle? Like what question did you try to get answered by that data? Because that becomes the problem, is people find data and they're like, okay, this is cool. Like in fact, people have told me about my example with the Massachusetts casinos that they're like, that's really cool but I would have never thought to do that. Like even if I was a casino customer and I found this casino data, all this data I would have never thought to do this, ask this research question and answer it. So my question to you is then, you know, like did you actually ask a research question and answer it because that's the next step is if you know, like I like to go to the casino, right? I like to play the slot machines but it's stupid to play the slot machines. You're gonna lose, right? So I should play the tables. I'm just like bad at poker. In any case, that's a bunch of subject matter knowledge right there that I just gave you about casinos. So I was able to take this data and do something with it. We'll look at revenue per slot machine and revenue per table. Let's see here. So long as you're talking about it, you said cheap as well as best wine available. So I'm thinking that the research questions that would be, here I'll make them as captions, okay? So what is the cheapest wine? That would be one. And then what was your other one was? Okay, two. What is the best wine available? Okay, so I'm interpreting your research questions. And I'm putting them here, right? I just made them as a captions. So if somebody brought me a wine data set and said, these are my research questions, I'd say. I'd give them a side eye and say, these are not good research questions, okay? And those of you who've taken my study design courses on LinkedIn Learning will recognize right away why I'm saying that. Because I give you like a formula for making. So the first thing in the formula is you have a sample in your data set usually. Usually you don't have a census. Usually you have a sample. And you've got to decide what is that a sample of? What are you gonna generalize to? So for example, let's say that I got the Kegel data set and I'm just pretending, okay? I'm just making this up. Let's say the Kegel data set was a list of wineries in California, like all the wineries in California. Like that's a census, right? So all the wineries in California. Let's pretend that was the data set, okay? Then if I looked at what is the cheapest wine available, let's pretend the data set had two tables. It had one table was the wineries and the other table was what wines they carry, like just, you know, so it and that's where the prices were. So I'd be like thinking, okay, I can take just the wine price table and figure out what is the cheapest but maybe one winery charges a different amount for the same wine, right? So I'd have to start, if you take that course, that study design, it's actually two course series, you'll see what I'm trying to do is operationalize the question. So I'd have to start like nailing down these details because this is how I'm going to generalize back. I'm gonna say, okay, the cheapest wine available in California from the wineries from the time of this data set was this wine from this winery or something. And then what is the best wine available? Well, you can already see what the problem is with best. What is best? And so you can't really use a word like best. I sometimes have to use a word like best because I do healthcare quality studies. So I'll say, you know, which is the best? Where's the best patient care? And I still have to operationalize it. And what I usually do is I come up with like a few different pieces of best. Like there might be five indexes for best and then I measure those. So that becomes the first problem. Let's see here. So number two is more generalized than number one, right? Is what Mohammed comments, I'll put it up there again. Yeah, generalized is not the word I would use but I kind of get what you're saying. It's more vague and vague is bad, you know, if you have to analyze data and answer a question because data is specific. So you wanna be more, a lot has to do with asking the research question because kind of like another course you could take from mine on LinkedIn learning is data curation. And I teach you how to make data dictionaries and find out what all your data mean, every single variable. And it's when you do, when I do that, when I curate my data that I start creating research question, I'm like, oh, you know, here we have a flag for whether or not this was imported. I start thinking, well, maybe imported is better than not imported or something like that. So yeah, lengths of taxification talks about good quality. So example, quality healthcare, quality wine. What is quality wine, right? You know, I like wine and I'll go buy a bottle and it'll have a little award on it because some judges judge it at high quality. How did they do that? Well, they had a few different categories that they rated and we do that in data science is like, okay, what's the best university? Well, we could take, you know, what has the best value for the tuition credit and put that in and put, you know, we could make a whole our own sort of like calculation, you know? So let's see your people, maybe people judge based best based on price. This is a good point. Like what do people want? And I bring this up when I lecture on ordinal variables, right? Like, let's say you and I go to the restaurant and you know, we have Yelp here, which is you can rate it one to five stars. You and I may think the restaurant is awesome, but I rated a five and you rated a four because you never rated anything five or something like that. And so that's the thing is like people just judge things differently too. So Spudny Dude says, I will go for now and rewatch later. Well, sounds great, but I really encourage you just pick a topic and try to ask a question about the topic. Like that's the best way to do it. And then you can do a portfolio project and then you have something to talk about if you ever get interviewed, right? So actually the topic of today was like how to research companies before you interview for a data science job there. Like, you know, like I worked at the army and one of the things I say in my blog post about this is public sector companies, you don't need to really research as much. You more need to research yourself to see if you're the kind of person that should work at those. And I definitely was. I was a government law for like so long. I loved working for the US government. The problem was I'm in public health and as you can see, we defunded public health in our government in the US so we're paying the price now with this prolonged COVID-19 pandemic. But if you're like, if you do cybersecurity and there's a field that doesn't care about getting breached, then you can't really do a job in that field. And so I had to change sectors. I had to go into business, you know, private sector. But still, you know, like I have a company, if you wanted to come work for my company, I would hope you would research it, right? Because then you would know whether you would get along with my company. I encourage you to go to my blog post where I show you an example of researching a company that has a data science position open that where you really just don't know much about the company. Like the company itself didn't even seem to have a webpage. So it's like, what am I getting into if I apply for them? And what I did was I just demonstrated on the webpage that you can actually gather a lot of information about these companies before you go there. And it's really important. And actually I'm gonna demonstrate now, unless anybody has any questions about, let's see here I'm gonna share. I'll just go back to sharing my screen here on, let's see. Right now, one says tax evasion says right now I learned how I learned how to clean the data and visualize the data. But I don't know how to make a prediction out of the data. Well, that's the problem is because you don't have a research question, right? Like actually let me show you something here. I'll go to my blog. And let me just show you the casinos thing, right? Just because it's easy. I made it for demonstration purposes. So it's so easy to use what I'm trying to make a point. Okay, so here's, let me make sure this is, okay so actually let me, so you can see the whole screen, okay? So I'm gonna sort of talk about it. Like down here, this is me talking about casinos, right? So that would be like, if you were writing a white paper or blog post, this would be you talking about wines and you were interested in the data sets and the points you were making. Why did you care about the cheapest one or whatever, right? So this is me explaining all that. Then here we've got a descriptive analysis. Like this is just how many slot machines there are, right? Like this is not like a predictive model. So that's the first thing you wanna do is, have I completed it? That's all I do is freelance projects completely. And so that's why I get my face in everybody's projects, right? So that's what Mohammed asked if I do any freelance projects but that's pretty much all I do. So that's where I'm getting all of my information. So back here, so here's the data. Now I want you to look at this, I did this in R. What is this? This is a time series, right? Like this is the year, okay? And what is this? This is a slot payment percent, okay? So what is the slot payment percent? Well, on average, we have laws that make it so, if I put money in a slot machine, it's not just gonna take my money and never pay me. It's gotta pay me something back. So what they do is they have this thing called RTP or return to player and it's always less than 100%. Now you want it to be as close to 100% as possible because the likelihood that you'll get your money back is higher but land-based casinos all over the US have pretty low ones like in the low 90s, okay? So what I did was I was able, they reported in this public data what their slot payment percent was. And as you can see, this is 100% and this is 85% because I was telling you like for example, online casinos have a much higher slot pay percent like in the 97% and where I am in Massachusetts, if there's no fun to go to the casino if they don't have good restaurants or good music or something, why not just stay home and gamble? So they're competing and I know this is true because I'm a customer and because I did search. So this very simple plot answers a question like this is Flane Ridge Park, it doesn't have as good, like it started terribly. And why are these different lengths is because Flane Ridge Park opened back here but MGM Springfield only opened here. And Encore, Encore is the one that's like a flagship when it's gorgeous. It's like a Las Vegas one. And so you can see here, it hadn't been open very long here. It shot up here right after it opened and then it went down here. And what is this? This is a control chart stuff. This is like a 90, I think it was 95% or 98, no, a 95% conference in a row or something I did to just say that the only outlier I really saw was there. But anyway, the reason why I'm going over this whole example is because my point is that you don't have to do, like you learned how to clean and visualize the data but you don't know how to make a prediction. I don't know how to make a prediction either but if you look at that, I can predict that if they don't get a better RTP, people are not gonna go to that casino, okay? You know what I mean? Like you don't really need, like I do multiple logistic regression. I do it like I did it yesterday, right? I'm good at fitting models but you just don't need a model most of the time. Like most of the questions people ask in data science can be answered with descriptive statistics like that time series. Oh, thank you. We have another person on the chat. Welcome, welcome. Yes, and oops, I didn't mean to put that up there. So I have, so go ahead and ask any questions and I'll answer them. The topic for today was originally me giving you advice on how I research companies in data science. One thing I didn't say in the blog post I wrote about it but I think it's good to do is to talk to other people like that you might know who are actually working there. And I used to do this a lot actually when I would work in the public sector. Like if I was working, like if you see my blog post the problem with working in the public sector especially in academia is there's just really bad management in a lot of places. I don't know why that is. Like I don't know why in government and in academia stuff there's just so much bad management and in nonprofits. I've worked for private companies and you think yeah sometimes there's bad management there. I've seen that too. But bad management that just goes on and on and on for years and years and years. I haven't seen that as much in private companies. So it's more when you're working in public sector and you're trying to get a better data science job it's really important to sort of roam around and find if other people are working on that team if they're doing well. Oh, so long to talk to the tape. Longs to talk to vacation. I need to talk to her. Ask her one pro tip for new learners like me. I think this is what I would say is a good pro tip is keep a lot of notes about each project you do. And by project I mean when you're in classes like if you take a course online or you're just in a degree program at college most of the time they'll give you homework but once in a while they'll ask you to do a little project. And also like something like the Massachusetts Casino project, take a lot of notes about what you did. Now you might be like well in word and I'm like well data curation. Take my data curation course and that tells you what notes to take. Make a data dictionary for all the data sets you use just at least for the variables you use and you make, you transform using your analysis. Like right now if anybody were to ask me about that casino thing I'd be like I don't know let me look it up and I'd have all my curation and I'd be able to get right back into it. And like take notes and even take notes about if you're in a group project take notes about who did what. You know, not in a mean way just to remember. That's what I would say. That's what everybody doesn't know at the beginning and realize it's way too late. And if you ever feel like you had a project where all you did was take notes and you almost did no programming, you're probably doing it right. Because if you take a lot of notes you don't have to program much because you know exactly what you're doing. You're not messing around with data very much. That's irony. So why do I do it? Because I'm a bad programmer. I'm not good at programming. I'm good at data programming now. I can't program a front end out of a paper bag. Like I'm just not good at programming. So that's why I started by taking a lot of notes. I'm a fast typist whatever. But in the end I realized that's what I wished I'd be more. So lungs detoxification is not a good question. Which is, is SQL important for our data science job? SQL is important for a data science life, okay. I will tell you, I've never had a job that required SQL. Yet, I learned SQL as soon as I can. So why would you do that? SQL is kind of like English. If you know English you can kind of get around everywhere in the world, okay. I'm sorry it's like that due to colonialism. But it's helpful if you're an English speaker then, right? Like that's why everybody tries to study English. Even if English is in your first language you can go pretty much anywhere and somebody speaks English, right. SQL is English in data science. You gotta know, everybody speaks SQL. And if you don't speak SQL you don't know what you're doing. So yeah, I would learn SQL because notice how I said everybody speaks SQL, I mean it. Like when I teach SAS I speak SQL. I say, let's say you wanna create table where such and such blah blah blah. Like we speak in SQL. And why do we speak in SQL? Think about what SQL stands for. Structured query language. It's a way of speaking about data. And if you ever take any of my sort of courses or whatever where I get into stuff about SQL. Like especially if you take my course and getting started with SAS on demand for academics which is the free online version of SAS I'm all excited about. If you take my free online course on how to get set up with that, I teach about SAS. I teach about the history of SAS. And I teach about what I call the SQL revolution which happened in the early 2000s where SQL was invented. And I say invented because it had been theoretically invented before then but we didn't have the computing power to actually create a query language where you just say in English what you want your results to be and an optimizer inside does artificial intelligence basically and serves you your results. Like we just, you know that was just amazing that you could do that. And the earliest SQLs were really the optimizer is really bad, you know so it took forever to get your results but now they're really good. But the reason why I teach about it in the SAS course is because SAS is not a declarative language. It's a procedural language and SQL is a declarative. Like you declare what you want and it figures it out for you. In SAS you're the optimizer. You have, you do what the optimizer doesn't SQL. So you have to know a lot about the data steps. That's the data step language of like that's you controlling the data in SAS. So that's why my intern hates SAS. That's why everybody knew it as data science is the one that learns SAS. But if you're in health, if you're going to health data science, you've got to learn SAS. Let's see here. So what are the essentials for a data science resume? Means what a data science job resume should have? Well, you're talking about resume but one of the things that a resume leads to is an interview usually. And so the resume has to say things that are gonna be brought up in the interview. So that's the most important thing is that you have stuff to talk about in the interview. And actually I was preparing a blog post. Keep track of my blog because I'm gonna post this blog post about interviewing. The hard part about interviewing, and I'm relating this to your resume, is that people are gonna ask you about being a data scientist. And if you haven't really done any data science jobs you don't really know what it's like. And so you don't really sound like you know anything. And that's the trick is that if you have done the data science job it's not so hard to be interviewed. But if you haven't, then you need to like simulate yourself having done a data science job so you can act like one. I mean, not like you're lying but you can just have the confidence of somebody who's done one. So that means basically trying to do projects it's better to do a lot of different projects that are small like that casino project, right? That casino project is a project. Let's say you were like me and you wanted to get in with casinos. Let's say you did five little projects like that with different people. Like maybe one with a professor you had or one with a different colleague. Like I did that one with my colleague I like to go to the casino with who's a fintech data analyst herself. Like a data scientist. And so we were like, let's see if we can attract the attention on the Massachusetts gaming commission. Maybe they'll want to contract with us. We can do some data analytics, we're big fans and we can give them some good advice, right? So like right now my intern, she's doing little projects. She's in her master's degree. She's doing little projects with like me and with her classmates, with other people. She's teaching with people who are working a lab. Like she got into some sort of lab and they had some data on molecular or something. And so doing these little projects is what you put on your resume. Like if you already have a data science job and you've done projects at that job then you put those on, right? Because what you're doing is you're queuing up what they're gonna ask you about at the interview. And you're also queuing up what you're gonna point out in your cover letter to them. Like, hey, I ran a data science warehouse, a SAS data warehouse at the Army. Like I could say that in the cover letter, if on my resume, that's what it says, right? But if I ever run a data warehouse, maybe I interned at the data warehouse, I could say when I interned at a data warehouse, I did the study on knee injury and you know, there was something like that, you know what I'm saying? And so that's the essential because you have to have projects that they can ask about in the interview that you can feel comfortable bragging about so that you sound good. So can Kegel provide me a job? So this is the way I see Kegel. I see Kegel as like, like if you buy rice and you see that as an unassembled dinner meal, all right, Kegel can help you but you have to do a lot to finish, right? So what can Kegel do? Kegel can provide you use cases which it's actually kind of hard to find when you're not in a data science job. So they can provide you use cases and they can provide you data that's well documented that you can answer questions with. The problem is that often stuff that's free is not that interesting. So the fact that Kegel's providing you these use cases in the state that I can answer them is a sign to me that these things aren't interesting because if they're interesting, then people don't give it away, right? So if you take the Kegel data and you take the use case and you can find a different question to answer with the Kegel data, that's been to me what I think I could use Kegel data for but the problem is I haven't been inspired. Like when I was a fashion designer, I used to go to Minnesota fabrics and I used to just walk through Minnesota fabrics and touch the fabric to get inspired. What I do now is I walk through the data warehouse and I curate the data to get inspired, okay? And I recommend that that's what you do with the Kegel data. So Muhammad asks, if we work on freelance project during years of employment gap, how can we highlight them on our resume? That's exactly what you do. Actually, I'll kind of describe. So people who do contract work and data science, they tend to have like either one or two page resumes, okay? And in the one page resume, the bottom half, like the top half is usually like their education and like special things about them like if they speak French or speaking English or whatever they do. And then the bottom half is changes and it has their most recent projects. So let's say that you were a freelancer and you did a project through ABC recruitment, right? That was the name of them. So ABC recruitment placed you at XYZ telecom and you did a telecom project. Then you say that. Then maybe, you know, ABC industries hired you to do an analysis. Like for example, I have a friend named Adrienne and she is the CEO of her company which is called Onyx Spectrum. Isn't that a nice name? So Onyx Spectrum is an IT company but every once in a while, they get a data analytics project and they've gotten a few data analytics products from transportation agencies at the government at different governments like state governments. And whenever Adrienne and Onyx Spectrum gets one of those, they call me and they say, Monica, can you do this data analytics thing for this transportation agency? And so now that's exactly how I'd put it on my resume if I was trying to get another. See, I'm not out there trying to get contracts, I just have this friend. Like I have some friends that once in a while they have a contract they call me. And so, but if we were trying to apply for contract or whatever, I would put through Onyx Spectrum, I did this thing for the ABC Port Authority or whatever it was. Let's see here. So, oh, LinkedIn user is back. I had to step away. Has anything been discussed about researching companies and data science so far? No. So please ask me a question. So I can answer it and be on the topic. How much a fresher data scientist earn in US? When you say fresher, I'll assume you mean junior. If you look at my blog post that's associated with this live stream that I put, that junior data science job opening that I highlighted was for about $75,000 to $100,000 a year is their annual payment. I believe that you could get that. I believe that if you're applying at a kind of crappy place that doesn't have any money they'd probably offer you 50 to 55,000 as a low bar but most people wouldn't take those jobs. I think probably most people, like if you get a master's degree from a data science place you probably shouldn't take a job for less than 75K. I kind of agree with that. And me, so the problem is if you're a woman or if you're black or brown, there's a lot of racism and they won't give you what you're worth which is why I started my own business. So if you keep finding that they won't even pay you the bottom, you kind of have to start your own business. But as a freelancer, that's kind of, I mean, I do freelance stuff but I just kind of have a business too. I also make curriculum and I teach and I do a lot of different things. But if all you do is freelance contracting for data science, you can make a lot of money. You can get good at that and just do that. And that's all you do. But I don't like doing that. I like working on a team but I don't like this team that team. I make a better leader than follower. And so I always have to be doing something more managerial to enjoy my, and that's another thing about researching companies is can you see yourself at that company? Because one of the problems is that new data scientists often are just like, I don't care where I work. I'm like, yeah, you do because it's so much work. Like here's one of the points I was gonna make is that interviewing for jobs is time consuming and expensive. So don't waste your time interviewing for companies that you know, once you got to know them, you won't wanna work for them. Like I'm vegetarian, I wouldn't work for a company that was big into meat, into the meat industry, right? And if something is called like ABC Industries, how do I know? Okay, so let's see here. You were mentioning a few minutes before the live started that it's good to make sure you actually enjoy the job. So some companies are popular, but the work may not be the best. Actually, that's true. So I will tell you what makes a data science job good or bad and it's really the manager. It's hard to manage a data science team. And I'm really, really good at it, but most people are terrible at it. Why am I good at it? I have no idea. I have really no idea why I'm good at it, okay? Like some people are just some natural talent. And then, you know, and my dad was a project manager. He's a mechanical engineer. He was a project manager at General Mills for years. He built all these factories and ran all these factories. Okay, so maybe it's just sort of genetic that I have all this leadership skill and I like running big projects and stuff. Most data science managers don't get there because they love managing. I don't know why, but I think it's because managers have a lot of power and they love power. I like having power, but not that much. Like I would rather just be happy. So I like being a manager to like make the team happy and we all work together and have fun or whatever. But yeah, so you can't know what the manager is gonna be like on your data science team. But you can research companies and you can learn about their values and their values trickle down. Like if you go to my company's webpage, you'll see that my company has a big thing about social responsibility. That's true. And like if you go to like, like I'm gonna record a video kind of about this topic. In the video, I'm gonna show how like if you go to Pfizer's webpage and you go to Liberty Mutual's webpage, Pfizer is like a pharmaceutical company, Liberty Mutual's an insurance company. Liberty Mutual seems to be more concerned about your health than Pfizer, you know? So that's why you should research companies. You just wanna make sure you know, you wanna bother with what you're getting into. Lungs and taxification, you said track. Are you asking how to get on a manager track? I'm just wondering like, or did you try to ask something and YouTube screwed up? And YouTube screwed up? Oh, okay, LinkedIn user has a question. There are jobs where I can sense you're supposed to work around the clock. There you go, you're smart. If I don't want this, then is data science for me? It matters where you work. It matters the values. Like I would not make people work around the clock. In fact, you know, it's so funny. I was just reading an article. It was about a company called Noah's Bagels. I'd never heard of it. And obviously it's a bagel place. I think it was today or yesterday. The entire group of people who work there came in and all resigned in mass. Why? They had had a beloved manager. She was unceremoniously fired. They had put complaints up to the top. The top wasn't listening. They all walked up. This could happen at a data science place. In fact, I was just reading and wired. Actually, I have it over here. I was reading and wired about Amazon's cybersecurity. Okay, I'm gonna read this from one of their articles about Amazon's cybersecurity here. So I'm sorry, I can kind of hold now. The team was leaderless again after less than a year. With chaos at the top, other senior staffers and managers would leave too, leaving the group unsettled and lacking institutional memory. Projects got derailed. And security would lose its top advocate in high-level meetings, former staffers say. The division's team would hunker down in silos, sometimes fighting amongst themselves and operating without a strategic vision. As a search dragged on, some staffers began to wonder why it was so hard to find a new chief. We couldn't find anybody for the longest time, says Hayden, that's one of the people they interviewed. I think word had gotten out that it wasn't an easy place to work in security. So you're right on there that you could, like if you got a data science job at Amazon, I don't know how it would go. But it was likely it would suck. Why? Because all we hear from people who work at Amazon is that it sucks. You rarely hear somebody, once in a while I hear somebody saying that they like that they're experiencing Amazon. So I'll be fair, once in a while I find that, but most of the time you hear people complaining. Oh, you're similar to me in terms of leading teams. So it's hard to, oh, you're hot, thank you. I gotta show, whenever you compliment me, and you'll get shown. So you say you're similar to me in terms of leading teams. Well, the good news is nobody wants to do leadership stuff. So if you're on a team, you probably have a bad leader. And I shouldn't say probably, but it's likely you don't have a leader or you have a bad leader. Like this Noah's Bagels, they just didn't have a leader. And so you can step up to the plate and practice your leadership skills. You probably, you may not get any credit for it. And so that's a problem when you're like a woman, you know, like me. I have led so many things, but I never get credit because what I was surprised to find. I mean, you think about, oh, subtle sexism. Oh, you don't really think about a woman as a data scientist, whatever. No, one of the reasons I left the army is I was afraid I was gonna get raped, okay? 25% of the women in the army are raped. So what is rape? Rape is a way of gaining power and, you know, and I was running a data center and people were threatening, okay? There were not many women around. There was not much protection. My, I was working at an institute that was openly misogynistic. Like a person working there who is really old, he's retired now. He came to me and told me that he said, be careful. This is a very dangerous place for women, okay? So when they started attacking me and trying to cancel my job and all that stuff, I just quit and said, this is too dangerous for me. You know, I need to work in a safe environment. So if you know where I live, you'll realize it's a castle. Like I'm the fifth floor of a castle. Like I'm that scared from that experience. So I would say that's really important. Like I recently, I gotta show you this. This is hilarious. I recently heard about this place called Databricks. Oh my God, is that the sexiest name that you've ever heard, Databricks, okay? So I'm gonna share my screen here. I'll share this. And I'm gonna show you Databricks, okay? Okay, so Databricks, here's Databricks official site, right? So this is, I don't know, I guess I clicked on some sort of ad. Okay, this is Databricks, which is I guess an analytics thing. I'm not really sure what it is. But I was like, oh, well, let's look at about us, their team, you know, like that's one of the things. Okay, this is not working out. Is I look at who, okay, this is definitely a fail if you have trouble, okay, there's a company about us, three years ago, okay, Board of Directors. So let's look at their Board of Directors. Now what Databricks does is it builds artificial intelligence. Now let's just look at it. Okay, see these people, right? Okay, now let's look at our founders. Okay, that was their Board of Directors. So they're gonna set a strategic. Okay, here, see that? Okay, now we're gonna look at this executive team. Okay, okay, so now we just looked at that. And do you think I'm gonna work there? There's like literally no women working there. There's almost no women. And there's nobody who, almost nobody who is not bright white colored face. So they're building artificial intelligence. Don't you think half the people scrolling through YouTube are women? Don't you think half the people buying stuff on Amazon are women? How are you gonna build artificial intelligence without any women? They're half the world, we're half the world. So one of the things is I would not, like invest in data pricks. Like what's wrong with them? I mean, obviously you're not gonna do a good job of making artificial intelligence if you, you know, if all I saw were black women there, I'd say the same thing. Okay, you can't have just this sliver of society building your artificial intelligence because it's all about classification. So long to talk to vacation ass. Does age matter in data science job recruitment? Well, it matters from the standpoint of, well, there's different kinds of ageism, right? So if you're young, the ageism is that you're stupid and you haven't done anything. But if you actually have done a million things, then you're kind of like a whiz kid. So if you're young and you wanna come off as a whiz kid, which is why I say do a bunch of projects or whatever, and that's how you wanna sort of frame your age. If you're middle age, like you're in your 30s or 40s, you know, I say that's middle age because I'm imagining people who are like maybe 60 or 70 being older age, you know, job seekers. But middle age job seekers have done jobs usually or done something. I mean, even if they were bringing up children, they did that, you know. So when you're more middle age, what you've got is subject matter expertise. And if you've been working in data science anywhere, you have experience. So you wanna highlight that. And that's not so much a problem when you're middle age. Like people tend to do the best getting their jobs when they're middle age. When people get older, they start to run into ageism. And on one hand, like, so the people who have complained to me about ageism have been old white guys, okay? I haven't had old white women do it or old black women or whatever. Like I haven't seen that, right? And so because of that, I have only those use cases. And in the old white guys that I looked into, what their problem was is they weren't updating their skills. They weren't doing new projects, you know? And they weren't, I mean, they take courses like in a new software, but they weren't just really throwing themselves in the new thing. And now that I think about it, through that person I mentioned earlier, Adrian, I, and also on LinkedIn, I've met some engineers that have changed jobs recently that are very gray-haired and they're just, they didn't have problems because it's really just saying young at heart or young at brain, doing new, just doing new things in data science and sounding young. Like a lot of people are surprised, like I'm almost 51 because I sound young, I guess. And also I have my hair, that's my other trick. But it's because I just keep up to date, like this Databricks, I went and looked them up and that's actually, to get close to, like we're in the last five minutes here, so I'll get close to closing this live stream. Thanks everybody for showing up. I'm sorry, I didn't really stay on the topic, but maybe that's the way I am. You can get me off topic really easily. But I got stuff to tell you. It's really important that you research your data science companies even if you aren't gonna work at them. You really wanna know what you're getting into. Like for example, let's say that you're a manager and you're thinking of making a big contract with SnowflakeDB because you want them to do your back ends in the cloud. Or like, what's that data robot? I love data robot. I don't know them, but they're this cool AI platform. I want an excuse to use them, make an AI algorithm. So I have this positive opinion about data robot because I've studied them and I've gone to some, like I went to a data science central presentation by the data robot guy and the data robot guy, one of the data robot guys. I'm sure they have women there too. But in any case, it's good to know about these companies and data science to just stay on top of your game and know what people are talking about. And one of the main things you wanna do, like I say on my blog post, is look up the recent news that they are in. Like if you look up recent news that Noah's Bagels is in or Amazon is in, you can learn a lot about the company and what you might think of working there. Like, I'll tell you, like for instance, SAS, SAS software in Cary, North Carolina has a reputation of being a really nice place to work. They're very employee centered. However, I don't know if that's gonna last because I feel like SAS has had kind of a free ride for the government for a long time. And I think that ride is not as easy anymore because just because of R and because of Python and because stuff isn't encroaching in SAS space. But we're still gonna need SAS, so that's why I have my book on data warehousing at SAS. All right, well, are there any last minute questions before I close the live stream? Thanks again for everybody showing up. Hopefully I can have more than five people this time. I'm doing a little bit better of advertising it. Well, you're welcome. I'm glad that you showed up. So everybody should, if anybody's watching this and got anything out of this live stream or this recording, because it's gonna be recording, I'd really appreciate it if you'd subscribe to my channel and you join these live streams. When will the next live be on? I'm planning to schedule one on Saturday to take place on Saturday, but please subscribe because I'm not sure, you know, like I got chaos life. So this is my plan, but please, you know, just stay tuned to make sure that I really do it that way. But I plan to schedule quite a few because I don't know why it took me so long to realize this. I realized I was repeating myself a lot when I would meet with people. Like I'd say, oh, I'll meet with you. I'll give you some information. Then I was like doing the same things over and over and I'm like, well, why don't I do a live stream and some blog posts where I'm sharing what, you know, for the wide audience, what I end up sharing one-on-one with people when I meet with them. And, you know, my customers and once in a while, you know, just somebody will contact me and say, Monica, I have a question. So I'll meet with them on Skype. And then I'm like, okay, I'm researching data companies for them and I'm like, well, I could just teach people how to do this and how important it is before you get a job or before you apply for a job to make sure that that's really where you wanna work. Because take it from me. If you get a job and you get settled in it and you're like, I cannot stand this job, it is a pain to then interview and try to get a different job. So, well, thanks again for being here, all of those who showed up. And now I think I am signing off.