 Hey everybody, I think we are live. This is Monika Wahee. I actually am going live five minutes early because I'm getting used to this restream software. I'm hopefully streaming on three different platforms right now, YouTube, LinkedIn, and Twitter. And hopefully, whoever's hearing me, hopefully you can join and you can chat. This is only the second time I've done this with this restream software. The last time I did it, the people who joined from LinkedIn, they had trouble chatting. The people who joined from YouTube, that worked. But not all of their messages came through, so I'm trying to get good at using technology. Good morning, spaghetti dude. Welcome. You are the first person to say hello to me on my stream. So you're letting me know that my stream is working a little bit. So in case you're just wondering, why am I here? My name is Monika Wahee, and I'm a data scientist. I say that so flippantly because data science is this relatively new term, and I'm a relatively old person. So I've been doing this a long time, long before the word data science were invented or whatever. But I now call myself a data scientist because I try to stay young with the kids and stuff. But really, I'm just an old epidemiologist, biostatistician, informaticist, management person, whatever, like the 500 things you have to learn how to do a programmer to be a data scientist. Oh, somebody agrees with me. He says, I know, I know. We have to know too many things in our field, right? And so the last time I did this, which was the first time I did this, so who knows how good it went, I'm recording these. And so if you follow me on YouTube, it's a good idea to follow me on YouTube. If you follow me on YouTube, what you get is a whole bunch of videos. Oh, yeah, yeah. So SpaghettiDude is saying this. I'm going to show it. Can I make a little consultation with you later? The answer is absolutely. What I do when I do a little consultation is we meet on Skype because I like Skype. I'm hard to use. I'm bad with technology. I probably shouldn't say that, but it's true. But if you insist on Zoom, I'll be on Zoom or whatever. And so SpaghettiDude, what I'm going to try to do is follow up with you, although I'm so new with this. I don't know how. What a good thing to do is maybe if you follow up with me, actually, I'm learning the software here. Let's see, or contact me at, oh, yeah. Free Skype consultation. OK. So I'm adding this because I don't have any moderators or anything like rich people I have, famous people. Like one of my favorite streamers to watch is this guy. His name is, they call him Roger. He's the big jackpot. He's always spending money at the casino and winning all these jackpots. And I always think like, what about regression to the mean? But it doesn't bother him. So if you want to ever just talk to me about data science or usually I end up talking about everything with you because I'm too talkative, I guess. I'm a little queacious. But people say I give good advice. Go ahead and contact me. And what we'll do is we'll schedule a time on Skype. Everybody's all over the world, including me. And so we'll find a time where we're both awake. Or I can get up early. Let's see here. So you can leave a, oh, OK. So spaghetti dude, you're assigned to email me. And when you email me, tell me that you're a spaghetti dude. You know, when you teach a class and someone, I said, OK, you guys have my phone number. Text me if there's an emergency. And I'm like, there's an emergency? Which one are you? I don't have even my address. All right, so for now, I'm listening to you live. Well, thank you, spaghetti dude. Let's hope more people join. It says zero on my dashboard. But I don't believe that because obviously you're here. So the last time I did this, so thank you. Yeah, I'll look for your contact. The last time I did this, I realized it would be helpful. I mean, of course, I was just testing the software. But I realized it would be helpful to have a topic. And so I thought, well, why don't I make a topic? And then, oh, what is today's topic? Thank you. I said, why don't I make a topic? And I actually put it on my blog. Let me, I don't know if you saw the announcement here. I'm going to just place this in the chat. This is if you want to go to the blog post. But here was the discussion topic, which is hiring a paid data science mentor. And I'll tell you a little bit of the backstory of why I'm saying that, because before that, the questions I've been getting are, Monica, I've taken every class. I know Python better than anybody. I know everything. But I can't get a job. And I also don't know how to, like if I did get a job, I wouldn't know how to use the Python. Like I obviously, I can take any course, but I can't apply anything. And I don't know, and I'm stuck. You know, like that basically there is self learners. And they're just kind of stuck, right? And also a lot of self learners don't have a lot of money. Like they can't just go and go, oh, I'll pay, you know, some executive data science program to teach me something for a few thousand bucks, right? So a lot of times the thought of like paying a mentor sounds like expensive, okay? In fact, when I started my business in 2012, if somebody had said to me, can you, can I pay you to be my mentor? I would be like, no, I don't know how to do anything. You know, so, but I'll tell you what happened. And I mentioned it in the blog post. When I started my business, I didn't have any customers. And I was focusing on, I came from academia. I was focusing on helping people in their master's degree and their PhD graduate. Because I'm an epidemiologist and a bio statistician. I got, you know, I'm like the drums in the base. And you can play an instrument, we can make music, right? Actually, my brother's a drummer, but anyway, that's the side. But so what was I doing? I was graduating a bunch of people, okay? And a lot of them were international people, right? So they were in Boston. Some of them weren't in Boston where I live. Some of them weren't all the place in the US. Some of them weren't in England. You know, I was starting a Skype sort of business. I was getting a lot of referrals. But I remember when the first people graduated, I was like so teary eyed. I was like, I'll never talk to you again. But they just called me the next week. I'm like, what are you calling me for? And they're like, well, I'm at work now. Now, what do I do? Why do my job? I'm like, I don't know. So I realized that, you know, doing your PhD is a lot like doing a job in a way because you're independent. Nobody's paying any attention to you. And so when I mentor people through their PhDs, I'm like, OK, we got a project plan. And I'm setting up these regular meetings or whatever. Well, I was just these people. I just went on autopilot. I'm like, all right, regular meetings. You don't bring me your work. I'll help you. So what started happening after that is people who are plateauing in data science started coming in. Because I started making my YouTube videos and stuff. And I was like, oh, I'm a data scientist, I guess. I call myself an epidemiologist. I'm dumb. I'm a data scientist now. But the reality is I'm not a normal epidemiologist. I have a huge background in informatics. Like I build databases and rent data warehouses. And you are not going to find an epidemiologist who does that. So I could easily slip into data science mode. And it was cool. But since these people had a relationship with me, they were already used to paying me. And then we just kind of kept the relationship. I felt very comfortable continuing to charge them and continuing to help them. And what happened is I just got really good at it. I just got really good at taking people who are not in a PhD program and not in a master's program, who are just trying to succeed. Like I remember one of my mentoring customers had a PhD in actuarial science, which to me is like so smart. But he wasn't getting anywhere at work. And actually, this was kind of a long time ago. A lot of the LinkedIn learning courses I ended up making were sort of based on trying to serve this customer and customers like him. Because he knew, oh my god, he's just such a good SQL programmer. And he's just brilliant. But he could not answer research questions. Like the workplace where he worked at had, they did some online service. But you had to apply and get approved for the service. And they were getting kind of pissed off because people were dropping off during the application. And then they couldn't get approved. And then they couldn't pay. So they were like, and I was like, well, a lot of people dropping off are probably people who realize they aren't going to get qualified. And so they just don't. But then how do you set up a study design to figure out who is dropping off and who is not? And I'm like, and then they came to me and they said, we want to do A-B testing. And if you've ever taken my A-B testing course, it's called Experimental Design on LinkedIn Learning. Honestly, a lot of that came out of me trying to set up A-B tests for this guy and a few other people where I'm just like, you don't need a test. You don't even know what's going on with your marketing. You don't even know how many people come to your website a month. So it was, oh, oh, here it sounds like you're saying something, but what's wrong with my chat? I can't see. OK, I would like to tell you that today I was actually thinking about this monitoring thing. I know it's a coincidence. Oh, yeah, you say this monitoring thing. What exactly do you mean like monitoring the rates, like the conversion rates? Because see, this is the thing. Now I'm confessing everything. I'm this epidemiologist. And epidemiologists are all about rates. Like if you tell me, oh, a million people died of COVID. I'm like, what's the denominator? We don't even care about numerators without a denominator. And so I'm glad that you totally understand me. And so what ended up happening is I started mentoring data scientists, like people who saw themselves as data scientists and people who are trying to succeed in positions that are called data science. And I started to realize, I think the problem is that it is really hard to succeed in this field. Because you really need to have a lot of leadership and management skills that usually aren't the type of skills that statisticians have. Like I would manage statisticians. I have a lot of leadership and management skills. And I would manage statisticians in my way of, like I always have this stereotype of statisticians and data analysts as these people who want to just put on their headphones, right? No meetings, just put on their headphones and program. Like that's the way I see them. And so, and not that there's anything wrong with that. It's just that they're not people who want to come to a lot of meetings. They may be social in their own way, but they're not like work social, okay? So I would have, sometimes you need people together at meetings to just make an agreement or just have everybody there for something. So I'd be so careful about bringing them to a meeting and then not having them at meetings but having data curation available to them. And if you don't know what that is, please take my course on LinkedIn Learning because it will solve your life. That course came mainly out of my mentoring. Just about every time somebody come crying to me, data 911, I'd start by making one of those diagrams that I show you how to make in that course. Yes, more skills beyond the technical things, thank you. Not only are you here spaghetti to do, but you're speaking, I'm so happy. Oh look, maybe somebody else joined. I see this changed to a number two. Yeah, that's the problem. It's not just like management skilled the way like you could get an MBA and then now you know how to do transformational management. If you just do that, it's not gonna enough. And how I know that actually I have a really good close customer who is a professor of business and she's really good and she's really smart but she couldn't be a data scientist. She couldn't do what I do, okay? She hires me to help her, you know? But I mean, she teaches statistics like she's really good. It's just that just training in business and knowing the management doesn't mean you can do it with data science. Oh, and there's another, yeah, that's right. There's another shy person. Oh, and then they left, right? So, but I tried to publicize this a little more. I don't know, you know, maybe next time. I'm gonna try and make do a nice job of posting these data science chats on my channel in a nice playlist that and I'm gonna try and make it so you can link to, we'll see, I have great plans. But anyway, you know, mainly because a lot of people are asking for advice and I thought I'd like pull stuff together and talk about it. So, what I talk about on my blog post, I learned how to share my screen. So, I'm gonna do that now here. Let's share the screen. This is the blog post. And then, so what I talk about on the blog post, let's see if I can coordinate myself, is, well, first of all, I'm cheap and so is everybody else, right? So, why pay? And I go through a lot in the blog post explaining how I figured out it's just a bad idea to not pay a mentor. Just don't have one if you're not gonna pay them. If there is somebody you want as your mentor and they don't wanna be paid, just see if you can arrange to pay them. But that's already a red flag because I'm happy to charge for my mentoring services because I know what I'm doing. I'm totally professional. But somebody's like, ooh, I don't know about charging. Well, maybe they're not that good of a mentor. I guess what I learned is that you can be a teacher and you can be an analyst, but being a mentor is sort of a set of skills. And you wanna be experienced before you start mentoring people you don't know or offering that. And so that's the first thing, is just throw away the idea that you should have an unpaid mentor. Look for a professional mentor who just does that. And people like me, we do more than just mentoring. Obviously, like I'll make curricula, I'll write papers, I do data analysis. So, but they, you know, which is good because then your mentor can teach you about real life because they're also doing real life. Oh, somebody did ask, why you deferring mentor from a teacher? Actually, that's a good point. And I kind of bring that up a little bit in my blog post. A mentor is kind of like, let me make an analogy. Let's say that something bad happens in your life and you're really sad. You could go to a therapist, right? Like a mental health therapist, or you could go to your friend, okay? Now, they're both gonna probably help you, but they have different roles in helping. So, if you go to a teacher, what is a teacher? Well, they're normally people who run classes, right? That's what I would say a teacher is. So, if I'm a professor, I'm a teacher. I'm going to a class and teaching a class. Or maybe if you join an online course, you could say Monica's the teacher of this LinkedIn learning course. But I see a teacher as linked with a curriculum. Like the teacher is delivering you this learning thing, a course or something, this information, and then you learn it. And they should be good at that. Good at saying, okay, you don't know R. Here's how you program R. Okay, now you can program R. The mentor is different. That's a different relationship. A mentor is somebody who listens. It's kind of like a therapist. Oh, hi, Kanak. I'm sorry, I didn't reply to your email. I just, I thought of it today before I started to. Here, let me look at, I should apply to it right now. Let me look at your email. Whoops. I probably shouldn't put it on the screen here. Let me look at your email. I'm sorry. Because you said, oh, you wanted my WhatsApp. I'll give you my WhatsApp. Actually, if you know what my phone number is, you can go to my webpage here. Go to my webpage and just look up my phone number. I don't wanna dox myself here. And that is, I'm in the U.S. And that is my WhatsApp number. So go ahead and connect with me on WhatsApp using that number. And I'm sorry, I'm really bad at using WhatsApp because I'm American and we don't really use it. So connect with me and tell me that you're at Kanak Sora and I'll accept your connection. And, oh yeah, and you wanna come to the U.S. So yeah, like definitely, we'll talk about you coming to the U.S. I love that. Yeah, it's been, hey, can I ask you, SpeedyDude, where are you from? What country are you in? You don't have to tell if you don't want to. But I'm so happy to be in Boston. I love being here, even though it's kind of cold. But I like cold because I'm from Minnesota, which is even colder than here. But yeah, but thanks for showing up, Kanak. And now there's three people here. I'm so happy. But feel free to jump in. What I'm jumping around, I never stay on topic. What I did was I created a discussion topic. Kanak can watch if I get any better at this. Which is like, should you hire a paid data science mentor? And not even should you, but what should you think about if you're gonna do it, right? So Kanak's in Bihar, India. That's, I don't know, it just sounds beautiful, right? Boston's beautiful in some way. So yeah, so one of the things that people kept asking me was, Monica, I don't know how to apply. Yeah, so I'm gonna continue the topic. So people were contacting me and saying, Monica, I've taken all these courses and either I just don't feel good. Like I don't feel like I can apply these. I don't feel like if I had a problem at work, I could do it. I could answer the question. And also they're also saying things like, I can't get a job, like I haven't applied it and they know it. And so I was like, okay, what's a good answer to that? One of the good answers is we'll do a project, right? And those of you who follow me, you'll know that I did an example project. I got some public data about the casinos in Massachusetts and analyzed it and made recommendations because those casinos, I don't like the, you know, I'm vegetarian. I guess they don't cater to vegetarians, but I'm trying to get the casinos in Massachusetts to listen to me and other ethnic minorities and just like make it, I don't know, just a better place for us. But anyway, I decided to use that opportunity. Since I want the casinos to hire me for data analytics, I thought we'll all use that to give me an example of a portfolio project. I would say a new data scientist could do if they were trying to get hired by the casinos. You're just an example. The problem was after I was done with that, I realized, well, that would be too hard for somebody who just took a bunch of my LinkedIn learning courses. Like for example, if you took my LinkedIn learning courses and study design and an RN and SAS, would you have been able to do that casino project, right? The casino project, I just downloaded data. Actually, I could even just show it to you. Let me see if I can find the, let's see here, casino. What I did was I just downloaded the data from the internet and analyzed it. And when you look at it, you'll be like, well, wait a second, I know survival analysis and this is probably the simplest thing I've ever seen here. Okay, so of course, I even made a little video about it, but there's only three casinos here in Massachusetts, so not a very big place, right? But if you look at it, the first thing I do is I just, these are the three, I'm just looking at how long they've been open, how big their gaming floor is, because they're gonna make more money, like as you can see, this one, MGM and Encore are almost the same size, but the problem is they're not making the same amount of money, like MGM is not making a lot of money. And you can see there's only two of them on each of these because there's a service that they have like this table games that's not everywhere, right? And again, you have not seen any statistics yet, okay? You've just seen me like doing the descriptives and kind of like curation, kind of getting to know three casinos. What can I tell these casinos that maybe they'll hire me, right? Well, I figured out that I could make a ratio of how much money they make per table and how much money they make per slot machine, because in their simple data, they said at the end of the each month how many slot machines they had on the floor and how much slot revenue they had and how many tables they had on the floor and how many table revenue they had. And in case you're wondering, the reason why this is all public is because our governments, our taxpayer-funded governments have supported these casinos to set them up. If they also give back to the community in different ways, and I do have to say Encore, which I go to a lot, is really beautiful outside. They really cleaned up a superfund site. So, I'm sort of partial to them, I like them. But if you can kind of look at this, this is my ratio here, this one is longer. The reason why it's longer is because this one has been opened the longest. This is Plain Ridge Park, it's not a very nice one. But then see here, since the green one opened, it's always been like lower. And this green one was the most supported by the government. Now, spaghetti dude asked, why did the government support them? Well, the green one I'll talk about specifically. MGM Springfield, so Springfield is a city that was kind of economically depressed. And about 10 years ago, I had a boyfriend, we thought, let's go visit Springfield. There was nothing open, it was garbage-y, it was like a ghost town. Okay, well, it's also a historic place. And so MGM went in and with the support of the government and the community, they really rebuilt the area, made it very beautiful, even moved some historic buildings, took the facades off of them and put them on the hotel, they made this beautiful casino. So that's kind of what's going on, is casinos just make money, you can't stop them from doing that. So our governments will try to get together and say, hey, in order to have a casino here, we have to make laws that allow it and regulate it. Are you guys gonna give us money? And then they set up that the casinos give back tax money, and that's why there's this reporting. But anyway, so I get a little off topic, but I can stop sharing this. Oh, let me just give you the link to this in case you're just curious. So that's an example of a portfolio project. Okay, so the bottom line is that is a descriptive analysis. I used ratios. Okay, that was simple. How do you know to do that? How did I know to do that, right? Well, I have an undergraduate degree in fashion design and textiles and clothing from the University of Minnesota. Okay, and you don't. So I guess maybe you need a mentor. So let's see here. So then, oh, hello, Gula, ma'am, firstly. Oh, thank you. Very nice thing here. So you're asking, so let me put this up, because if you say good things about me, I'll put you up here. So you're a big fan of my team, totally love it. Thank you. Can you please guide me towards being an NLP engineer? And so you mean natural language processing, right? Because NLP can stand for other things. You're saying currently you're doing a master's in science and applied data science and statistics. Okay, so at the University of Exeter UK, okay, so just a second here. When I tutor mentor people in the UK in the graduate area, it's different than the US. They don't help you, okay? Not that the US is awesome, but US colleges feel obligated. They make their professors meet with students and at least pretend they're helping them, but not in the UK. So whenever I've had anybody studying in the UK, it always is kind of expensive because I end up meeting with them a lot. I'm basically, what I end up doing is I end up trying to figure out what their project, what rules their project needs to meet and who needs to sign off on stuff. And then I try to design a project and I keep having the learner go back to them and say, is this okay? And then once they sign off on something, then I just work with the learner and we get that project done and they get the heck out. Like I'm like, nobody's teaching you, you know? Literally, I've had learners since 2012 and there's only one who's gone to a UK college and learned, she actually learned a lot, but it was London School of Economics, Middle Eastern History Program, so that's not data science, you know? So I strongly recommend the LSE, Middle Eastern History Program, but I don't recommend any school in the UK, okay? So if you're doing an MSc in Applied Data Science and Statistics at the University of Exeter in UK and you want to become an NLP engineer and you mean natural language processing, which I'm assuming you're doing. So one of the things, it depends on what you want to, so if you read my data science, newbie starter kit, you'll realize that the first thing I make you do when you pick becoming a data scientist is pick a topic and so you're like NLP engineer, what am I gonna, immediately I'm forcing you into a topic. I have a new, relatively new, she's not new anymore, she just graduated, I had this customer in linguistics and she gathered a corpus, I was like gross, like a corpus, no, it's not a corpus, it's a corpus. A corpus is where there's a whole bunch of talking and it's written down and you have to analyze it, I'm like that's not, how do you do that? Well, that's what you do when you're an NLP engineer is you go okay, so there's all this talking, like how do I make it data? So there's these things called tokens, which is like where you take a piece of it and you say okay, this is a data token, right? My customer was in linguistics and she was studying noun phrases. She was studying like the status of noun phrases. So she went through and coded this corpus so she got these students to write some stuff and she coded these noun phrases as tokens. So she circled all the tokens and then she coded them as to linguistic features, right? So what NLP engineers try to do is see if they can do that in an automated way. And so there's a center in Boston called Maverick that it's an acronym. It's part of the Veterans Administration System. And I knew a guy there, he doesn't work there anymore, but he's really cool. His name is Leonard Devolio, you can follow him, he's a smart data scientist, he's really cool. But anyway, I remember over there he had a data system and people were trying to do natural language processing on reports, like if you have a tumor and the pathologist cuts it up and writes about like cancer or not cancer, that's all natural language. And so they were gonna try and see if they could classify these reports using natural language processing. And what the basically is you end up in a world of validation, right? So here's a hundred reports that I know it's a positive for cancer, some of that I know they're not cause I can read it, I'm smart, you know. And then you design Goulomb, you design something that would have separated these automatically, you figure it out, right? So it's too hard, right? But sometimes, so in 2000 people, I would say, oh, give me a break, it's too hard. In 2010, I was like, oh my God, that's too hard. In 2015, I was like, natural language processing is useless, too hard. Now you can do it, it's not too hard anymore. It's, you can do it in chatbots, you can do it, like it's useful now, like there's use cases, I just love it. Oh, thanks, Spaddy dude. Yes, definitely write to me and be as nice as, what's this face card of who I didn't get back to and then I see him and it's so embarrassing, one person shows up and I didn't get back to him. So yeah, definitely write to me and tell me your Spaddy dude when you write to me if you don't sign up with that. So are there any advanced projects for natural language processing? Well, that's the problem. The tips and tricks, you have to tell me what is your field? Cause I just gave you two use cases. I gave you one in linguistics, right? And if I had, like let's say I met with you, Goulomb, and I taught you about the linguistics project I did with this lady, and I taught you about the challenges and the issues we had, you could then go away and come back and propose and a more automated solution to our challenge. And so if I was your professor at the university, whatever, and I'm working with my colleague on linguistics, I'd be like, oh, colleague, help me work with Goulomb. Let's see if we can improve your experience of assembling your corpus with natural language processing. And now we have a use case and now you have a project and now you can graduate, right? Or if you're in Boston and Leonard was still over at Maverick, well, I don't know who's at Maverick now, but I'd be like going saying, well, you want to be in healthcare data science, let's go over and start trying to find a better way to classify these reports. And see what's important, Goulomb, is I didn't know anything about linguistics when I met this customer about a year ago. And I told her, she was like, help me graduate. And I go, well, help me learn linguistics. I don't know what any of these words mean. So token, I knew what token meant because I already knew a little bit about NLP, but noun phrase, I didn't know what noun phrase was. And I didn't know, like, she was studying target-like articles, okay? Have you noticed that when people speak English and they have an accent, they don't always pronounce the words exactly the way like I do, where I don't really have an accent, okay? So like, you know, like sometimes, like I've noticed people who speak Spanish, they might say more than we do, you know, people who don't speak Spanish at all and just speaking English in the US, right? So that thing where they say, that's considered not target-like for the US dialect, right? So if you ever hear like, like a person with a little Spanish accent, you're like, oh, that's not target-like. That's what the term is. Okay, I didn't know that term. So the problem is with natural language processing or anything you do, survival analysis, whatever I do, there's this whole set of business rules in that topic that you don't know the terminology, you don't even know. Like I didn't even know that people would care if something was target-like or not. Or I didn't even know that you have, that that's an attribute, right? People who are new to healthcare data science are really often very surprised about all the crap we have in healthcare. Like they sort of imagine surgery and things, but we have a lot of crap in healthcare. Like I was teaching at, I was doing some seminars at the healthcare quality department and we were talking about nosocomial infections, preventing hospital acquired infections. And I said, what is a good way to do that? And everybody said together, use a bundle. And I said, use a bundle? I don't know what a bundle is. I've been doing this for years, but I still don't know everything, right? So somebody told me what a bundle was. I don't even remember what it was, but you see what I'm saying is that, that's why in natural language processing, especially because you're dealing with like words, you have to pick a topic first. So that, I would say here, I'm definitely gonna, let me see if I can find my, I'm gonna look for my, this blog post I can give you glum because it kind of like lays out a data science journey. And it's actually linked to a video I did. Maybe it's better if you watch the video because the video is more focused on like just generic, not health data science, although I'm partial to health data science. So here's a link I'm putting it in the chat. And actually I'm gonna, I'll share the screen here. So this is what I gave him the link to. This is the video. I really think you should think about it because, so this is what I say in here. I say people who are coming to me and asking me, how do I do this journey? There's two kinds of people. There's people who already have done some sort of career, which was I guess kind of me. And then there's, and like that actuary I was talking about, and then there's people who have not like they've, they're like new, you know, they're newly going to college. They've never had a job before. If you've never had a job before, it's harder to pick a topic because that's the first thing you have to do is pick a topic. So I remember a long time ago, I met this guy who was a real estate agent. This is a long, long time ago. It was like in the early 2000s. He was a real estate agent. I met him at a party and he was telling me how he got kind of sick of selling houses. But he realized there weren't any really good apps for keeping track of the accounting for selling houses. So he made an app for that. And then he started selling and then he started doing that instead. And so that's a really good example of how it's better to start out as a subject matter expert and then become a data scientist than that. And I give you tips actually in the video if you're beginning your career and you really haven't worked anywhere before on how to find out if maybe you already know a lot about a topic. So you can start with some knowledge. So you can, like one of the problems you're probably gonna have go along is like how do you do a project if you don't have any knowledge of the business rules or how that topic works? And so you kind of have to pick a topic first. And then the next thing you wanna do is you wanna learn about the data and the measurements in that field. So I just gave you an example of like this linguist, they're worried about noun phrases and target like articles. Then I gave you an example of these people at the Veterans Administration who are worried about cancer pathology reports. So that's the problem is that if you're doing NLP it's gonna be totally different whether you're doing it in linguistics or you're doing it in this healthcare setting. And by the way, you can work at the VA and do this cancer thing. And then your next project is brand new. It's on cardiovascular disease and you don't know anything. It's like you have to start over. But at least if you're still talking about veterans and still talking about healthcare, it's not so bad. Like that's, I ran a data warehouse at the Army for a short time, like just a few years. But literally it was like that for me because I knew nothing about the military. Nothing, okay? And you can't do anything with their data unless you understand their terminology and their business rules and everything. So I was like, oh my gosh, I'm gonna be useless at this job. Well, it turns out I'm kind of a quick study. And I just talked to soldiers, there's soldiers everywhere. They like service people, they explain everything to me. And so I quickly, I adopted the vernacular and everything, but that was step one is I had to just figure out what was going on because the Army is not just some normal old system like accounts receivable department or something. Like it's got its own business rules. So even though I was good at like using data and data warehouses to answer epidemiologic questions, I could not even answer like one question until I interviewed everybody to death. I didn't know where our data was coming from. So why is it that that's that way? It's because data are basically measurements. They're measurements of things. So if I like when people say this car gets so many kilometers per hour, it doesn't register in my head because I'm in the US and we use miles, right? So if I like don't really have a feel for kilometers, that's useless data to me. So the most important thing is to understand your measurements which is why I'm so into data curation. If anybody's taken my course on LinkedIn learning, it's because anybody can curate the data. You don't have to even know how to do data analysis. You just need to know how to ask people questions and write down documentation, which I thought was obvious, but it's not really that easy. So that's why I made a whole course in it. And that's why, you know, if I'm ever your paid mentor, that's one of the things we're gonna be doing is a lot of data curation because and I guess Gulaam, you say tips and tricks. This is a generic tip and trick, you know, it's kind of like get vaccinated. You know, if you wanna get, if you wanna inoculate yourself, so to speak, against data confusion, use data curation. I mean, and like, you might be like, oh, well, what is data curation? I mean, I kind of have some blog posts about it, but it's basically making metadata that's not necessarily just, you know, like there's data dictionaries, everything is data dictionaries. You may be heard of that, but it's like making diagrams and making other kinds of documentation in order to keep track of what happened in your project. Let's see here. I'm looking for anything on, like a good blog post, my data curation. How about, let's see here. No, let's see here. No, I can't find it. Well, let me put the link to my course in here. Whoops. I just keep a, I have this list of links. I'll put the link to the course in case you're interested on LinkedIn learning. It's called data curation foundations. And I'm literally the only person who teaches it because I didn't know what it was called, okay? And so when I was at the Army, I think it was at the Army, somebody said, oh, you do data curation. I'm like, what's that? Cause actually, if you look at my CV, you know, I used to be a fashion designer. And one of my things I did was I was an intern at the Smithsonian Institution in Washington, D.C. And I did curation of clothing. I did clothing curation, like historical clothing. So when they were saying I was doing data curation, I'm like, you mean like the museum, you know? And then it kind of made sense, right? And why did I fall into data curation? I'll just be honest with you. I can program and obviously I can teach you how to program. I can do good job programming, but I program like a manager. I program to get the project done. I don't program very efficiently. I'm not very good at programming. I don't go fast. I don't program a lot. You know, I program whenever I need to and then I forget how to do it again. And so I'm not like the best programmer, okay? But what I can do is I'm like a producer. I can put it all together. Like if I can program something, I hire somebody or I get somebody to do it, you know? And so that's, this idea that data science is all programming is like kind of bad, right? Because like, I just don't do a lot of programming. And I know people who are way more powerful than me in data science, who do no programming, right? And it's because it's actually not that hard to learn programming. I'm telling you this and I'm not that good at it. But like I can do it, you know? And there are people who are just like really good at it, you know? And so what's hard is the stuff that I do that no one else seems to be able to do. Like data curation apparently. You know, what's hard is putting it all together. Like one of the things, like the theme, I was, a few more people have joined. I'm gonna say this, this is the theme here. Was why would you ever hire a paid data science mentor? And the answer I would say is it's so hard to impart knowledge to people unless you have a use case. And you wanna have a realistic use case. And so it's really helpful to me if somebody gets a job and we meet like once a week and they're like, okay, this week, our team is supposed to be doing this, I don't know what to do. And you know, I sign an NDA with them, like I don't share anything with people because I'm from health here, you know? I'm all high book applying. And I just try, I teach them curation. I help them curate their data. I help them figure out what's going on, you know, in there and then figure out what to do. And then they go back to work and they look like a star. They look like they're so smart, right? Yeah, I mean that costs money. But then if you get promoted, like if you immediately get promoted pretty soon you're a manager and stuff, like it kind of paid for itself. So, and I can't guarantee you that will happen because some people work at dysfunctional places. And I'll tell you that, I'll tell you if you won't succeed at the place you're at. You know, like if you know anything about me, you know, it's like I'm honest, I'm not gonna just make stuff up. So, and also if somebody comes to me and says, will you mentor me in data science? And I don't think I can be helpful. I'll just say no, you know, I don't just take people's money. But yeah, if I think I can set up a program for you and I can actually help you, you know, I would do that. And also, like I say in the blog post about it, which I put in the link, maybe, you know, my skills are not the ones you're trying to get or my mentoring isn't what you need. Maybe you need it from other people. But you know, like for example, I don't really do a lot of machine learning. Like I mean, if somebody has an assignment in it or, or if somebody hired me and said, Monica, I just love you, help me learn, help me figure out how to optimize this machine learning algorithm. You know, like people do that all the time. They hire me to help them with just the craziest stuff. Like I've never heard, one of my interior design students, yes, I have interior design students because I'm fashion designer, remember? She had this cool application. Her assignment was like to design an interior for this apartment, a specific apartment building in New York City. And she was given this application and it had, I think it was a graphic database. It had in it all the specifications for windows and light that can go through it. And all the, and all the knowledge about the Latin longitude and the light hitting New York City and the times. And what we had to do was simulate stuff. We had to sort of select what windows we wanted and then simulate things and write stuff down and make recommendations. I mean, unfortunately, the student I was working with is like a total like artist. Like she's not into data science. But she was like, oh my God, you know, and so, but I was like, oh, look at this, this is so cool. Look at that, look at this, Monica, what are we doing? And so, you know, I'm like really good at like just kind of riffing it, just kind of figuring it out. In fact, a lot of my customers have told me that what is best about me is not necessarily data science. It's that I'm good at taking a problem they have and just kind of breaking it down into something we can finish and just, you know, it's good enough, I'll say it's good enough for government work because I use work with government. You know, just so we can get the project done, you can get whatever you're trying to do done. This is if you're stuck in the middle of something, you know, I'm good at like getting you through. In fact, I wrote something on LinkedIn about it, you know, because a lot of times people will be, you know, in a situation where they spent a lot of money and they spend a lot of time on something, either getting a degree or trying to get a job or something, and then they're like, oh, now I'll spend more time and money on you, right? And I always say to them, you know, I'll meet with you for free and make a plan and tell you how much it'll cost. And it's up to you if you want to do the plan, you know? Sometimes people don't want to do the plan. But usually what is happening is they don't want to do the plan and they do something else and then they come back to me after a while and they're like, okay, now we have to do the plan in a hurry because the other thing didn't work. I'm like, why don't you just do the plan, you know? Because sometimes people don't know what they want, like we, I've helped so many people get degrees that they don't want the degree when they're done, you know? They didn't realize public health was this or data science was that. There's so many times people want to study something and then we're done and they don't like it anymore. Not all the time, just once in a while. And so it's kind of like, even in, like I have long-term relationships and mentors do, but you can have periods of mentoring where you just do one thing and you get through it. I recently posted on LinkedIn about a use case where I was helping one of my customers with health quality KPI thing. And she had been a tutoring customer before, but then I didn't hear from her and then she was at work and they had to do this KPI measurement thing. And she contacted me and I mentored her through that project. And so it's kind of like, that's why you would want to pay a data science mentor is to help you apply the knowledge that you have to real cases, especially if you're at work. So you succeed at work, basically, that's it. And, you know, work is not just data science, right? It's interpersonal stuff and now it's like COVID-19, you know, there's a lot of stress at work, right? And there's just a lot of things that happen at work and it depends on the workplace. And so I end up doing like a lot of data science therapy is what I call it, is where, you know, data is something that makes people fight. And the reason why people fight about data, there's two main reasons. One is that a lot of people think if you have data or you have access to data, you have a lot of power. So they fight over that and I don't fight over that. Like, cause I usually have access to all the data just as a practical manner. And so I'm like, I don't think it's very powerful because my opinion is to get access to data, you just usually have to apply, you have to go through some approval process. Like, you know, you don't have to own all of it but they fight over that because it's a power thing. And then they also fight over it, people who aren't into the power thing, they end up fighting over it because they just don't have a shared understanding of it. You know, they just don't have a, like I remember, I used to be married a long time ago and my husband worked at this place where, it was an engineering place where they'd build these big machines and the sales reps would create a quote and they'd mess around with the quote. And then when they decided to build it, it would become a job and they'd copy it to this other database. Well, the problem was when it became a job, people in the engineering shop would change some of the specifications because they'd realize they had to, you know, because these are custom built machines, they're like big food processing machines. And my husband was like, my ex was like, they fight over the fields. Like the sales rep wants to look up the old quote and see what he quoted but the engineer has already changed the quote because he's doing the job and I'm like, just take a copy of the final quote and put it in a job's database. Like just, you've got, like there was something before quote, there was like, like, spec and then quote and I'm like, just make a new thing called job and make that a copy that engineering can change the copy. And my husband like ran the database. Like he could put tables in it, you know? And he was like, I just remember he screwed up his mind. Are you sure that, like positive that'll work, you know? And it did. And then sales stopped fighting with engineering over this. It was that simple. Like they're all, this was like a family business and my husband had been running the database. He'd built their like 10, 15 years and they'd been having this problem. It was really that simple, but that's the issues. This is more of a design management. Like how do you come up with some simple answer like that, you know? You know, obviously this guy, all these smart people I've been thinking, but one of the things that was about that workplace, it was all men because they, you know, if women tried to apply there, they'd laugh and throw the resume away. This is what my husband used to tell me. So it was all men working there except we're like secretaries and stuff. And then also they were all engineers. And so if you get, everybody is the same. You get this group thing. These are just bad businesses. This is like we work. This is like, you know what I mean? Like these places, like you'll see on my blog, I post about Theranos, you know? As a bunch of group think of stupid stuff, you know, who thought that you can take one drop of blood and do even two tests on it that are totally different. Like it, and anybody who's worked, knows anything about path labs, knows that that's not how they work. So, but anyway, I digress. So hopefully I've answered some questions so far that people have, because there's not many people on the stream. Hopefully I'm anticipating good questions about paid mentoring for data science. I have seen different mentors out there that I think you can pay for data science. And actually, if you want referrals, I probably shouldn't just talk about people in general. But if you want referrals, write to me. Like if you say, well, Monica, you don't teach what I need to know, but this is what I need to know. Like then I would, but you really need to tell me your situation because if you're at a workplace, then that's the best time to get mentoring. Is if you're an, and actually, even if you're an intern, and I hate to tell you this, but internships in general are not very helpful. And the reason is people who run internships at these offices are not good at running internships. I mean, they're good at their jobs. And so once in a while, you get somebody who's really good at it, but usually the school wants you to do certain things and do certain reports and things and have certain experiences of your internship and your school is a school and everybody's doing an internship at a different place. And so how do they get all these experiences? It ends up being kind of a mess to try and fulfill all the internship requirements, get some useful project out of it, get paid attention to by the place because they're a regular business, they're busy working. And so when it first happened, I thought it was weird that somebody would say, hey, Monica, can you mentor me through my internship, my eight week internship? Can we meet? And you mentor me through it. I was like, that's not supposed to happen. That's a really good idea. If you end up in an internship where you're like, I don't even know what's going on. Like they put me in a cubicle and they gave me a manual, you know? Like what do I do? I'm not gonna be able to do my paperwork at the end of this. I'm not learning data science or public health or whatever I'm trying to learn. You know what I mean? Then you definitely want to contact me because I have, like I'll give you an example. I was a tutoring student and she went into health quality and she was getting a master's degree in like health administration. And she wanted to do health quality. And I said, you're not gonna get assigned logical internship because you're going to a crappy university that just has a crappy program. And she's like, well, let's cross our fingers. So she was assigned to this weird place. She was assigned to a business unit in a healthcare system. She was assigned to try and figure out this problem they had had for two years. And the problem had to do with this. There were these services that were given some lung cancer screening or something that was done and the clinic could recoup the funding from the government if they uploaded a data set to a portal. But no one had ever been able to successfully upload that data set to that portal. I guess it's like utilization data or something without getting any errors. So for two years they'd been occurring these records that were not ever being paid for by the government because they couldn't figure out how to get the data into this portal. And I said to her, I'm like, okay, I'm a data scientist so I have bad news for you. How you do that is you read this documentation from like 1969 and Courier type and you have to format your data set according to these super complicated specifications. But then it'll work. She's so smart. She literally did it. She not only figured out the problem specification, she read them all. She wrote a manual. She got all that in there and everybody in that internship, they put interns through there all the time. Nobody ever saw that problem. She did it, she got the reimbursement. But you can see how if I hadn't been there, that would have been a totally wasted internship. How is she ever gonna figure out how to solve that problem? Like that had nothing to do with healthcare quality, right? So I'm just giving you this example because this happens in data science. Like unfortunately a lot of data science master's degree programs really are kind of haphazard. Like, oh, learn statistics. Now we're learning Python. Now we're learning, I don't know about management or something. I will actually recommend one. Boston University, the Questrom School has a data science program where I talk to the guy who runs it. And he's slowly expanding it. And it's fast, you go fast, but you get applied knowledge, you know? I think in Boston University, I always cost a lot of money, but I would say consistently, not just that program, but every program about BU. BU is not my favorite place in the world. I could make a list of a hundred things that make me angry about BU. But I have to say that if you come to Boston and you wanna pay, and when you're done, you wanna make sure you've learned and you have a quality education. If you're in Boston, that's Boston University. If you're in Minnesota, that's the University of Minnesota. It's way the hell cheaper, way the hell cheaper. So if you can put up with making snowmen and getting shot by police, then I would strongly recommend you go to the University of Minnesota where I went, because literally after hanging out in Boston for a while, where we're supposed to be all into colleges, Harvard is good if you wanna meet people and be rich because you were introduced to the right people. And the same with a lot of those other colleges. But if you really just wanna gain skills and just be good at what you do, I say Boston University, any other programs are good. And that's the way schools tend to be. They tend to be either all pretty good or all not that good. That's what I found from tutoring, is that although I will make an exception, Bentley is a school that's a really good school for business. Their data science program is bad, it's bad. If you want to learn about data, at Bentley just take a business program and take like Informatics, take some program in actual business that they've been doing for a while, do not take their data science program. So unlike BU, Bentley just hired a bunch of smart people and put them in their data science program. That's not really what you need to do. Like you have to make sure like BU does that you've got the right people to get the right experiences for the people in the program. So they learn not only programming, but how to apply it and how to be on a team and all that. Like, at Bentley, they just hired a bunch of divas. Oh, I made this machine learning thing. Oh, I made this algorithm over here. But those people don't know how to work on a team. I used to work for them, I'll guarantee that. But anyway, so I see there's only two people watching. I was hoping there'd be more people today. But I think at some point we had five people, maybe if I start doing this more often, people will come more often. But thank you to everybody who showed up. Please make sure to contact me if you want to ask me questions. You can contact me on LinkedIn. Also you can contact me by email. And I'm on social media. But anyway, like what I'll do if you contact me and you want a consultation is that we can meet on Skype. And you can just ask me questions and I'll give you answers as best as I can. Sometimes people end up using some of my resources or using some of my services, but a lot of times they don't. It's just that I haven't had any new customers lately. Most of my customers are referrals from other customers. So I just thought, well, I need to just start getting new customers, so I better meet some people. So I met you today, Golan. Thank you for thanking me. And hopefully if I do this more regularly, people will catch on and be able to come and ask their questions. Let's see here. I don't know why that. So as we're closing today, today's discussion topic was hiring a paid data science mentor and why you might want to do that. And I was just explaining, sort of the big picture is if you're having trouble in your job and you need advice, like you already have the job, or if you're just having trouble applying your skills in data science, you wanna do a portfolio project or you wanna do something and you just don't know where to start. You have skills, but you don't have to use them together, that's when it's really useful to find a data science mentor, meet with them, find out what you guys can do together and come up with a payment, a budget and a process and a plan and some goals and then start your journey with the mentor. And then you don't have to be alone anymore. And you don't have to struggle to code alone anymore. Won't that be nice, right? Well, thanks everybody who showed up and I really appreciate it. And to this data science chat and hopefully I'll see you again. Have a good week and a good weekend.