Breaking into Data Science w a Data Scientist from Delta Airlines

okay hi everyone uh welcome to the breaking into data session i'm aaron feliz founder and ceo of promotable with us today we have saharan and siri i'm totally sorry if i butchered your name um but um really cool background um so you guys are going to learn a lot really just about the data science job process i know especially if you're just jumping in it sounds like you know it feels kind of crazy you're getting questions that are left field you're not really sure what to do how you should approach it um so we'll talk about that and and more for that just a couple of quick announcements um be on the lookout for emails for our next upcoming events um next week we have a data science lead from milliman he's really cool he's also looking for interns so if someone is trying to jump in really great guys got experience a bunch of different bigger companies as well and then we have a data science recruiter the week after that we have two of those coming up and we also have a data scientist from amazon um towards the end of the month um so those will be really great um and then just a quick note about promotable and who we are um so promotable we're a data science career accelerator so we focus on closing that data skills gap we do that through a variety of things first and foremost through our events community so we have stuff like this almost every single day it's always free and this is just so that you know it's one thing to say i want to do data science it's something very different to talk to people who do it every single day so you know what you're actually going to be doing you'll know what kinds of different titles you should look at uh what it's like to work at different kinds of companies different scales all those kinds of things and so we cover a lot of that but really i want to make sure that everyone feels comfortable raising your hand asking questions you know everyone's here to help you and we want to make sure you're able to get as much out of it as possible and then for folks who want to dive deeper we also have a immersive data science bootcamp accelerator course which takes you from you know no technical background all the way up to data science junior data science level and we tack on of course all the career services and mentorship and network that we think are missing from a lot of the boot camps that's also an income sugar back program so if anyone's interested feel free to reach out to me after the after the event but for now um sahar in this area really really cool person amazing background please kick it off um thank you for the introduction appreciate it and it's happy to be here and um people if you guys want to say your background and shout it out in the comment section or chat section that'd be great so i could uh know what is what is the people's background here that'd be awesome um yeah feel free uh mostly we just want to know if everyone here is an aspiring data scientist you're looking to break in or if there's some folks who are already experienced feel free to throw it in chat section and then we're going to have a little conversation here for the next 10 15 minutes or so feel free to get your questions ready throw those in the q a uh we'll address those as soon as we we get to it and then also for anyone who wants to we always encourage you raise your hand join the conversation i'd love to see all of your faces um definitely please um please raise your hand as we get to q a um we'll we'll um we'll we'll bring you on because i think you know that's one of the things that we're missing from all these all these webinars right is being able to really kind of see each other ask those kinds of questions and get stuff from it so i think the first thing would be just if you wouldn't mind just giving everyone a brief intro on uh who are what you do what your kind of road was up until uh until you got to delta um and then and then we'll jump into it yeah absolutely um so i started my um college with us with industrial engineering and uh i actually chose industrial engineering because i had no clue what i want to do uh in my life and the thing is that um about my country is that people usually we have mandatory courses only mandatory courses we we don't have music courses or art courses back in call in high school so usually people don't know what they want to do for the future they don't know what they want to do um industrial engineering sounded really awesome to me because i thought um again we call it like a an ocean with one inch depth because you learn a lot of stuff but not very deeply um so starting with industrial engineering i felt like um let's figure out what i want to do i thought i'm more into fashion or i'm more into um i don't know like interior design but i don't know i ended up with um i ended up in robotics um and after six months i kind of learned how to do coding not really but i tried i started with c i bought i bothered a lot of people um to learn how to pro how to do how to do programming our job was with java um and then i remember one of my classmates told me that bill gates said um in 20 years people will understand the value of python because i was really angry that we have to learn python it's um at the institution that i was doing robotic stuff and then um and then but that was that was like a that was like an important thing it changed my mind okay if that's a thing i should i should get more into it and now um i have some skills but i'm very very confident about my python and then for grad school i wanted to i wanted to use all the things that i learned or collected as my background i didn't want to just be a computer scientist so um what i did i thought okay let's just find something that it's like all of them together and data science was uh was it was a good option and um i also work in a branding and marketing company and they um they told me that um i wanted to go to germany but then they told me like i'm a united states kind of person which i don't know if it's a good thing or bad thing but that kind of like get into my head again and i i ended up in the united states studying data science and then it was a year and a half it was master's degree um i don't know if it's a good news or bad news to you people but data science you can barely find a good job in data science no you can't find it at an analytics job but data science job you literally cannot find jobs unless you have that grad degree sadly because now it's going to work towards phd and i really don't want to do phd but um but hopefully it has not still started but a lot of companies like google they don't they don't accept anyone with less than phd unless you have five years experience um and yeah anyway so i ended up in the grad school of data science and then after that i started my job at delta march 2020 four days before lockdown which was challenging everyone on linkedin now is their link um no i mean really cool i think it's it's a unique story but also a common story at least part of it right it's uh especially there's a lot of folks who are kind of coming in with like master's degrees like how do i target it what do i do um and so it's super interesting but i think my first question really is just diving into the pre-delta part right talk to me about what that job switch process was like when you first finished i mean it was at the university of rochester right um so like what was that what was that like at first did you just spread out applications and like what was the response and kind of how did you optimize what what did you learn that they don't tell you when you first go through these programs uh so in our school they used to tell us like yeah just networking is very important which which is true it is very important but they used to say like a networking is very important and just like chat with people go to on linkedin or networking sessions and just like ask them how they like their company they're not going to answer you if you do that if you need referral just go and get your referral i'm working and i know that i don't have time to you know like ramble on about what's going on at delta it's just um i mean if they really want to know that's awesome but if they don't want to know i know that it's not going to be a really um valuable discussion if you i mean they're not going to really pay attention if you just want to know what's what's going on in their company just to get referral and people that sense it and it's like and it is right to sense it because you put like hey i'm looking for a job and then um i'm looking for a job and then you say okay i want to learn about your company i saw this position so there is some way they're like okay you just need a referral so what i do always on linkedin i'm just like send me your resume i'm just gonna refer you and they're like oh thank you so much so you can just like very this is what i did this is what i learned from school and i saw that it's not working um let's start from the beginning i started applying for job in august 2019 i was supposed to get graduated and get graduated in um december 2019 and i started from august and uh we had to go to this festival it's like a women in technology celebration of grace hopper and it was a huge thing it was a it's it was way over my head the size of the career fair was huge and all the big companies small companies like there were i don't know maybe 1 000 companies i just don't remember the pages but i remember that it was six pages of just the logos of the companies that they're going to be there as a in a carrier fair so it was a huge huge venue and you couldn't see the end of it and um it's so i i had to get ready i had to choose which company i want to go to and i'm not going to be able to go to all of them that's for sure because it was three days um three day celebration so i had to like sort them out what when which one except international students which one sponsor international students so there were like the things that was important to me so i had to talk to people online i had to get them referral um so that's because that's just for women in technology i'm not going to really go deep into that but that's why i started from august but then after that in october which was the after like at the beginning of the october was the end of that celebration i got a lot of um a lot of interviews in the in this in the festival there but i blew them because um some of them were i was overqualified because they were like data analytics and they were like business more business stuff um which they were like we're not gonna do python like you you're going to be overqualified you're not we're not going to be able to offer you the what you want um but some of them for example like capital one i i was i wasn't good i wasn't good in the interview because i i had a lack of knowledge and and statistics and i'm really lucky that i have i got my delta job because they didn't ask any statistic question i was really lucky and that that was the first time that i actually i was like okay so after i find a job i'm gonna just be focused i'm just gonna focus on the status again that's actually what i did i'm i'm i'm not perfect on statistics but i'm really good now um so i understand the concept i know the application anyway but i started with the um cap i think capital one was the longest interview that i went i went to the last round and it was so so hard to focus it was really really hard to get the questions when you don't know what they're talking about and i just wanted to i like for example i'm gonna give i don't i don't know if people mention their background but um for example like they asked me about boosting and i just i i memorized what boosting is and then in that pressured i forgot what boosting is and so i couldn't answer that it's a recon you cannot memorize things and that was my approach to memorize the the things that i have lack of knowledge in and then it didn't really help me i couldn't remember um but i noted the things that i need to i mean i have the notes that i have for my interview always you need to get prepared there that's that's a rule and um and then i yeah i did i did interview with paypal i did interview with um capital one microsoft blooming uh no not blooming girls bloomberg what is that um was it like the business research company yeah yeah yeah yeah bloomberg um they asked me actually software engineering questions uh i've no i couldn't ibm was ibm was like the ho like all of the people who did ibm interview in the in our the friend circle at neurovar they were like what the hell like it was so hard it was three sections and each section was like a really long job it was easy to understand what question is but it will take a while for you to finish it and it was only 140 minutes time for you to do it so that was like and then i when i saw them they were like yeah because we don't accept entry level and i was entry-level um it's it's interesting you brought up a couple things that i i want to touch on because i think they're like really valid and important things um like how much of it is as you think companies not really understanding not how to hire but like how to describe the roles that they're they're offering and how valuable do you think it would be um you know rather than like reaching out to people and looking for a referral but more saying like hey i understand that oftentimes like data science titles often sound similar but the work you do is is very different i'm kind of curious you know is this more for a statistician or a software engineer right it's like that's crazy but i get that because there's a lot of folks and even um or people who tuned in to last week's event with with scott who's now at northwestern mutual um he was working at a previous company who was like i joined as the data scientist but i ended up doing a lot of data engineering um and that wasn't in the job description yeah that's true actually my call i really at work i really insist on the project that they like i insist on the data science of the projects that they um they assigned to me because i'm not gonna do anything else and this is i love my job i'm not just i'm not just here to earn money and um i want to learn and so i but my colleague she is she she's from um she did environmental engineering and then she went to stanford and did a business school something and then in the business school she took uh i think machine learning course she's amazing don't get me wrong but sometimes she she's like oh my job is not really data science i'm just like coding she doesn't insist like sometimes you have to go and like hey my boss hey boss i'm this is this is not data science i'm not here to code i'm here to do data science you know yeah no part of it is like sticking up for yourself or not sticking up yourself but advocating for what you're interested in and also maybe trying to get projects that aren't assigned to you right there is that like i think i asked you earlier about working with stakeholders and there's two kinds of stakeholders right there's like your business stakeholders but often it's like making friends with people within your your group to kind of get the kinds of projects and really understand like what's value-add and and what's not uh in terms of like for the products you're on um but i'm also curious um like after going through all those interviews is there like did they feel like all different or like looking back are there a few things that you're like you really have to be able to nail this for most uh data science uh interviews that you go for um can you elaborate a little bit i'm not sure yeah is it like do you have to have like really strong stats background for every interview or it depends on the world um data science uh it's a general yeah but if your data analyst no it's just like be more business person or to do more visualization they care about how how you're going to present your work they're not going to care about the modeling that you're going to use i'm just i'm exaggerating of course there is like a good person bad you know it's just like it's not always 0 and 100 i'm just like saying the data analyst mostly is trying to explain the data data scientists job is to be able to model the data and then the engineering part they work with database and uh how to um they do more software engineering things they do um they might do like even uh optimization like code optimization make it better faster yeah data scientists they were just like hey like i we check and like random forest works but like the data engineer might even [Music] develop the random forest themselves to make it faster than python library itself you know yeah yeah that totally makes sense i think the key takeaway here is um do your homework before you actually go into an interview and read them not just the title right because usually like in the description if it's like we're looking for someone who's going to be making tablet dashboards all day long but they call it data scientists um that's what they're going to care about but um i guess my other question really is like what are there any takeaways from like now that you're actually doing data science work at a real company are there things where you're like oh man this is what matters in the real world um and like you know they don't actually cover this stuff in in grad school sorry i was thinking about the thing that i have to add for data scientists i was just thinking about it for data scientist interviews yeah they don't they say oh just no machine learning no python guys machine learning they mean statistics if you are i saw that some uh i want uh i list up to now one of the people there for he yeah he or she is in phd in statistics good for you like they find job so quickly it's amazing they know statistics that's what they need i'm i'm trying to learn statistic as well as a a phd in a statistics and it's so hard it's really hard to work and study um um you know simultaneously yeah that's great uh one of the things where yeah everyone goes through a boot camp does it it's totally insane you have to be a little bit crazy um i say that but also like when you're hiring someone you're like oh this person works really hard like that's what you want they want it um but one of the things you also said that um was interesting for me was you said you know the one thing you don't necessarily want to do is just try and memorize everything um but we're interviews what what do you mean by not memorizing stuff is it just like you know you kind of understand the patterns that things go into and it's more about solving problems or is it something else yeah just like for example if you just google what boosting is they're gonna just give you one sentence and you can just memorize it and say oh this is boosting and if you remember then cool because yeah they're gonna be happy about it because they're not but like they're like you're gonna forget because there are tons of algorithms how many of them do you want to memorize yeah really definition yeah and how important is it um it's one thing to be able to do the stats and make the algos but um being able to explain it to like a business stakeholder like why you use one algo versus another one how significant is that for um at least in your experience um for data science um i think we were talking about this in the uh backstage that um so if in terms of when when you find a job and you are a data scientist and a company um this is the problem that also my colleagues my data scientists felt they also have this issue that it's they're gonna present it for very old experienced people in that industry and for example in delta they might be like not old as the age but they're like 30 years in l airline industry they are they are the grandfather of you know grand grandfather of airlines like they know a lot so they're not gonna be like they don't care about like how you're doing things and they don't really enjoy like hearing fancy words and they might think like oh you're just like using fancy words to convince them and this is the atmosphere and they don't like it's not like oh i'm also like okay i'm going to talk about an example i had an ex i had this presentation i put in from my from my perspective was really understandable it was so good and i also added the pca part just in the y axis of my graphs that this is the values and this is the pca values that's it and my my boss actually but he wasn't the business level i was like i was actually looking for him to say like oh good job and then i have to go to business level and convince them he was what is pca and i started like explaining that this is what we do with pc and say i don't understand why we're using it he was frustrated he was like what the hell do you mean by this like i i'm not getting what you were doing here and i kept like saying it was it was making sense for me but he was getting madder and mad at me because i was just like you know using fancy words in front of him that he just doubts and don't know and he's like hey don't waste my time i'm i'm the director of this department i don't have time for this you know jibber jabber i want to i want to learn what you're what what value are you going to bring to the table that's what what i want to know yeah so yeah it's just like and for example again like we have no values in our data set and that's exactly the discussion that i have with my boss today and tomorrow i have a presentation from my director and i was talking to my boss that hey like how would i tell them that i because so the reality is that my my graph should have like should have like gaps it should be empty should be no value and now it's all of them are valued they have they are non-na value and my boss will say like hey how do is not this is not reality how did you do that with what with what rationality did you feel these guys yeah then i just tell him like oh pca algorithm in python doesn't accept null value he's gonna like what what are you talking about like i don't care like uh tell me what was your how is this gonna help me out you know just yeah i i i'm scared i don't know how to say that tomorrow so i ended up with a i just checked with my boss that i'm going to say like hey it's just a mathematic rationale behind it needs us to feel these gaps like you have to choose your words the way that the shell holders will buy it otherwise they're just going to be frustrated because just imagine yourself if you go to i don't know if you go to a conference that they use like very technical they say like oh it's going to be for any level of people of understanding of environment and then they started talking about technical things they're going to get frustrated you're going to leave their room they're i'm not going to understand what's going on so that that was the case with the shareholders i mean we don't have like a shareholders like clients but we do have like business level that we need to sell to them like hey this what we're doing is way better than what you're doing manually oh yeah i know what you're talking about is more common than than most people would expect i i was having a conversation with someone i can't name this company because i'll throw them on the bus but we'll say it's a big uh west coast tech company uh do with that what you will um but here's like at this company and mind you this is like you know one of the titans of tech is like they say that data science has no roi um and so much of it is because these like people who are doing it don't explain like what business value they're generating so someone just thinks we're wasting millions of dollars for nothing but if they say hey actually we're looking at these kinds of clients or we're looking at we know this if we can figure this out we generate way more revenue but if you can't have that conversation like he's like i won't join a project if i can't if if like they haven't articulated the roi or they'll get do something else because yeah they'll all be valuable um so i think that takes actually the longest time i know that the people and i mean that's a saying in data science community that would say like 80 of data science job is just data cleaning and that's true it's pre-processing and that is very true i think 80 percent of data astronauts job is to be able to sell their project because people don't know what they're talking about among those 20 left eighty percent of it is cleaning it's just like it's really hard to keep get people on board yeah i mean we're gonna have a bigger company like i was talking to one of my other colleagues and he's like you know it's if you're a startup you're making the you know the mli algorithms yourself but like if you're at a huge company like i am we use data robot or some other vendor because honestly they could do it better when you can and the biggest problem i have now is figuring out which model to use making sure the data that's fed into is correct and actually pitching that to stakeholders um i'm gonna dive into a couple of these questions that we're getting um one of the questions um and from emilio he just asked if they're like what resources did you use uh preferably free resources online that you use to prepare for your interviews um and i mean i know a lot of it obviously is for you know kind of if you do it you up and you learn um but there's other stuff that you you use to help yourself prepare um so here's the thing for interviews it's just like you're not going to have time to watch videos it's just like time is very tight you you want to get the best out of it you want to be able to cover everything you want to just also review the things that you already know and make sure that you you are there um some some of the is just um oh we have a we have a and sorry you are people here hey math phd nice anyway sorry um uh what i was saying i forgot okay now so it's just like your time is tight you gotta go to um you gotta go to your interview you're not gonna have like three months to even three months is not enough to go through everything i think the best thing is that um first get the guideline from the um the hiring manager like hey do you have any cool what kind of question like for example in capital one they were like hey i just want to give you a heads up that it's gonna be gradient descents gonna be one of the questions so we're just like getting idea like okay how deep they're gonna go into machine learning you know just like oh it's a gradient descent it's very basic so they're not going to go crazy deep into learning but that's a sad thing because i blew the interview that wasn't really deep so i mean sometimes interviews it just honestly depends on who you're talking to like there's even variability even like what they care about or what they think is valuable like i've heard some people who say like i could train you to become a better coder like you know um what i can't train you on is like being able to hustle or having domain knowledge to belong right so it's like depending on the company sometimes like what i tell people is like you're trying to break into data science and you're transitioning from a different kind of career but hey if you have like a financial background go to finance companies right it's like i have at least like you're gonna at least get why they're doing data science versus some other folks um let's see let's grab some of their questions here oops data science no towards data science oh no it's it's a good source it's not perfect because people just put things there's not checking there's no checking as long as much as i know maybe i'm wrong oh yeah tbs yeah and please correct me if i'm wrong but um that's what i'm under what my understanding is but just like try to search and quickly collect your um collect your data i put it on paper and i cop i print it out to have it you know on my hands all the time and i review them because and before interview you only have time to memorize that's what i'm saying like now that i have time yeah i'm studying statistics to get the fundamentals better so if you have time if you're not in job hunting situation right now it's a good time to collect what you want to do it's what you want to learn and um if that's the questions let's i do coursera courses sometimes i i have this um element of learning statistical learning i think it's a very problem i'm gonna send the um send the link in a minute but that that is a really great book i'm i'm actually reading it right now um i'm feeling that i need to go to bayesian instead i mean i'm not just feeling i'm sure that i have to go and touch the bases that sticks a little bit more deep into markov chain um and then corsair actually has really good courses i think from santa monica university i'm not quite sure i can share that link here as well emilio and um what else i hope is the best kaggle is the best i did one project i mean after school one project in kaggle last weekend i was bored and i was like okay i don't have anything to do um let's just do some kaggle and i started i loved it and you learn a lot because you got to search but it just like don't like the project could be very easy it's just like a simple classification but i ask a lot of questions for myself for example i ask myself how do i know that all the attributes here is related to the classification maybe they're not helpful i remember that i heard something in statistics that oh like there is a p-value for regression that you can just say like how valuable each of how effective each of the um attributes are i know there are all other like feature importance methods and stuff i i didn't want to use the libraries i wanted to get deep into a statistical perspective of the modeling that i was doing so i it took me like a project that i could do in 15 minutes took me two days i was just thinking about it thinking and asking myself statistical questions that might come up in your job in your interview you know that's i think that or those are the great sources also also before going to interviews you can you can contact people there data scientists there that's what i do too on linkedin and hey i have interview and once you say you have interview they're going to be super super supportive what do you think i should focus on i did this for coinbase they said hey you should focus on a b testing i don't have a b testing experience come on and i didn't get a job because it was a b testing interview you know i think it's one of those things where but that's also the way to go right it's like be strategic ask questions um and and honestly i mean i think if you're not reaching out to people on linkedin all the time who are currently data scientists like don't ask for a referral don't give them a resume but it's like hey i'd love to just learn about what you do or like i'm doing research to make sure i'm like preparing my interviews what are like some common interview questions that i should prepare for right yeah ask me some advice but don't ask for anything else but exactly and glassdoor also has a lot of really good research for example like big companies they have their um they have like their um medium.com uh articles that they say what they're gonna ask like people put like i put like oh amazon frequently questioned for you know data science role and like it's going to be tons of tons of information i did the same thing for delta and i did the same thing for uber but um but before going to the next question i also want to say that for referrals what i did i used to get up in the morning at 200 300 people i'm not joking i'm not exaggerating 200 i'm i'm very consistent when i decide to do something i'm very consistent so i used to add like four and sorry 300 200 people each day um yeah my i think my uh before looking for jobs my um my connection with the number of connections that i had on linkedin was 282 something and then after two weeks i had 2 000 people connected yeah and literally i used to when so i used to add them in the morning and then go to class coming back go to my network and see which one of them accepted my request and if they did i'm just going to type them hey this is a horror i'm graduating from this very short they're not going to have time to read the whole thing i'm reading um i i really want to go to this company this position is open and i was wondering if you could do me a favor and look into my resume and if you like it you could refer me they're usually ninety percent of time they're going to say yes they're really really really helpful and the other thing that i what i did when this was my approach in october but i got better at it once i was doing it i realized that oh instead of like adding 200 people a day or 300 people a day let's be more um specific to the things that you want to do i was like okay i find a job posting i'm not gonna apply for it i'm just gonna find the data scientist people in that company and connect with them and i say hey i find this job i'm really interested can you refer me that was another thing and that actually that got me way better jobs than way better referrals and jobs than my last approach i hope that i asked the questions i mean yeah honestly it's just but it's just putting yourself out there and like the more consistent you are that it's funny you mentioned that because i did the same thing when i first like was looking for my you know last time uh when i moved companies and when i i live in chicago now when i first moved to chicago um i never understood that and it took me a while but then same thing 100 200 people a day and it is so shocking how many people who are willing to give you a full hour of their time i had a tech wonder give me a full hour of billable time for no reason telling me which vc's to get reach out to which tech accelerators to go find a job at like yeah people feel like you're watching they're going to be very generous with you if they're not expected right you get what you get i mean not everyone's nice but if you do 100 a day and you're specific with your ask um you know if you're like hey just want to learn more about this and you're not taking anything from them you know a threat to them most people are willing to help you but if you if you just like spray your resume out there like you're going to be like me i get it all the time and i'm just like delete it there's nothing i can do with this right like i don't know you it's extra work for me uh i don't really care um but if you can like prove that you've done some effort and say like oh this is my background i'm you know just about to finish up i'm interested in this kind of stuff i'd love to learn more that will pay dividends um hard right you're gonna do it when you first do it feels like you've sent out like a thousand emails and then you have one conversation and then you learn the jargon you learn the words you learn like what seems more valuable and then you refine and this is a very data science thing right you optimize yeah job search this is how you know i'm a nerd it's all it's a thank you and i think because it's it's a little bit controversial but uh because it's a good thing i'm just gonna say i think it's a cultural thing but i uh or because they they this nationally like these nationalities they have more um uh international people in the united states i got the best referrals from uh people uh asian people i don't know their nationality but they were asian people and they are so generous and i got from other people too i'm not just saying that but they i think they're because of their culture i thought um i kept asking them like hey what is what is in your culture that you guys are so generous i don't know my chinese friend had told me that it is in their culture to help other people because they were in a communist country for a long time and the focus is group not individual that's like i don't know i'm i'm very thankful of them and whenever i see uh see them i ask them like hey this culture is amazing let's promote it to other nationalist definitely no ways complaints and we'll take who we can get um but also i think part of it is right i think also like as a i mean we haven't really talked too much about like this element of it but like it's like you're an international student right i think if you find other people who have also gone through that process they know how much of a hassle it is and that's that's being being uh being nice it's like a pain in the ass youtube's gonna reject this video but that's okay um it's i'll just re-upload it like i usually do um is what it is um but you know it's like one of those things when you go through that process right you know that like like i know when i was hiring i was like look i love you but like my boss is not gonna give me 10 grand for a lottery ticket to attempt to hire someone and like learn their process especially a smaller company bigger companies they have old apartments but even then if you're neck-and-neck with someone else they're like i don't have to go through the uncertainty any extra cost but when you have someone who's done it they get it and yeah like finding people who get it whether it's that or something else you have some other kind of connection um i mean especially when you're trying to get that first intro um you know if use what you have if you're an international student go by international students if you me went to ucla go hit up all the ruins you can find even if you're not super passionate i can't talk about my own nationality our persian don't help each other that's a thing it's not really helpful i mean i'm sure that people wouldn't get offended okay look they're laughing that's true that's a thing and we know that so um my american friends they don't really understand the stress yeah but i got the best best help from asian people and oh my gosh i love how helpful they are and they like literally when i say like thank you like too much they're like oh you don't need to thank me i'm like are you kidding me like i'm so grateful yeah leg up yeah no exactly um what about um there's a lot of folks here who are like various questions from michael scott someone either yeah it sounds like two first names i'm sorry if i'm insulting you um but we'll get to michael's questions um so michael um if you want to just throw them a chat because you say you're a beginner kind of person looking into getting a getting data science are you going through a data science boot camp are you an undergrad um that'll help us actually answer your questions i think you're curious about what tips you have for your undergrad that helps if you read this thing um so michael's undergrad what should he be learning to break into data science as a for your first job i think um being among data scientists is good um i mean online coursing is amazing but um i'm that's that's my style um that i want to be i want to be there in person and i think that's that's that makes me to learn better um sorry i'm just like sharing the link that i promised this is the book that i told you guys um so yeah it's just like it helps me to excuse me sorry no no one wants to hear me sneeze it's loud and frightening so we're all happy it's you instead um yeah so yeah it's just like um it's um and what about um for learning uh if you want to dive deeper into python obviously there's like data camp and that sort of thing um the biggest complaints that i get about that don't rock start there um but you want to make sure as i had a head of training for big financial company talk to me and he said we had our traders go and learn go to a data science track on data camp but the problem is like it's their own proprietary environment and these guys didn't know how to spit up a jupiter notebook afterwards um and so he was like it's kind of it was a waste of time for them to do that and not get the practical stuff so just make sure like do it to learn scripting do it to learn like the basics but then go into kaggle or somewhere else you know my master's degree with the background of like i i did a lot of python like i knew what machine learning is and i think um i mean i don't know again like the background of all the people here but i know that data scientists they all know android g i think i took the course years ago yeah and um it helped me a lot and it's actually that's that's um and that's going to be yeah yeah that's going to be a perfect start um i learned a lot um and uh yeah i'm sorry i'm just trying to find the resources for people as well as i'm talking no no no no no worries i mean i think the resources honestly if you haven't been on the kaggle yet that's the first place you go um there's lots of good um there's like contests there there's really crazy smart people but you can go through lots of data sets right everything from like bike sharing data to airbnb stuff to don't do this while you're hungry like i did ramen shops in new york um you know probably the best choice but really fun stuff and you get actually see what's there obviously the difference between that and the real world is those are all really clean data sets and they're all there for you so the real world is like as as um that's actually one question that i forgot to ask you what's what's like the um kind of like the difference between doing data science and like an academic setting right versus like stuff i'm gonna answer that but are we gonna finish um that michael's questions i just don't want him to wait too long um yeah which is kind of your question too what is the best way to learn um so i did python in it it's not very popular i don't know why but it's the very good website uh learning website team treehouse uh it's um i don't know i learned python from there and it was really good for me it worked really well for me um after a while you can you can see that oh you know the rest of it yourself you just don't need to pay anymore um yeah well the monthly ones are nice in that respect yeah so i share the the link with you guys that's where i learned but i'm sure that there are tons that you can do and there's nothing always there there's nothing better than get your hands dirty with god and as you mentioned also kaggle like you can start i don't know how basic you are but michael but i think you can um it's a good thing to if you are not like super beginner you can go to kaggle but if you're a beginner i think any course would help you out to start yeah i mean yeah any of like the online ones code academy data camp whatever it just gets you there thinking like one do you like it right this is like the most important thing if you hate python data science is not the place you should be probably yeah people like put focus on python because python is very complicated but guys don't over estimate the underestimate actually the value of that's sql the biggest oh i need to know that every director that i talk to they always complain that the boot camps over emphasize python and then whenever we hire people and they suck at sql and they come from a lot so um that's like actually i would say before python sql is what you need to learn at least from my conversations and i'm glad you agree uh and then and then python's next um and then honestly one thing that also comes up at least for my conversations is soft skills um being able to get the stakeholder being able to just talk about it you know be that bridge between the super technical and like the not crazy technical yeah that's true and i i'm going to go to your question as well but just like yeah but like if you know if you don't know python you're not going to be able to be a data scientist but if you don't know sql you can't even enter to that world like data if you don't know python you can at least be data analytics but if you don't know sql you're not going to be able to even be data analytics that's that's the thing yeah i mean there's like no code tools but you're just really limited if all you can do is unless you're an amazing database person um you know if you have someone else like make the dashboard for you if someone else query your data um yeah yeah and and um but they don't know what you're asking for like i'm i'm bad at people i wouldn't call myself good but i know enough to say like okay well i'm gonna go have someone else do this for me i need to understand what i'm asking for because one you'll have a lot more respect um but also when someone says like so i get a lot of um like manager types you know marketing or product what have you and they're like what class should i take to understand like i don't want to do this stuff but i have people on my team who do it i need to know if it's yeah project or a 10 minute project yeah make sense to answer your question and also michael um every my everyday life and the projects that i do i usually work with sql daily to craft data depends on the project like i have two projects one of them is totally completely based on python i just collect data from sql it's that that was the easy part that i have done it like sorry i did a long time ago and my focus is the modeling and generate the visualization get it ready for stakeholders having it on the presentation those are the things that i'm focusing on and another side project that i have which is the beginning of the optimization and data science project but the thing is that to begin it they need to put what manual into automation and that's the hard part it's very hard codeine based job and i hate it it's with sql it doesn't have data science i keep saying like oh i don't want to do it but that's that's the thing if you're data scientist sometimes the project is not ready for you to jump into it you have to make your project ready for yourself and that's what i'm doing right now well that's a huge a huge thing that we don't talk about enough is like the data science uh workflow if you will like most people think oh we're just going to be doing this modeling or everything machine learning things but if your data is not ready like if you don't have a good question right if you don't have it you don't have the metrics you're going for if you haven't like struck you know uh looking something up we have this but we need to get all those kinds of things like it's a lot more complicated um and oftentimes those are the decisions um before anything else um like for the modeling part i i remember that when i was doing uh like for example pca and my job i was getting a really weird um information for my pca there where where it should go is that the the graph where it should have gone up it would come down and my boss was like what do you think that the problem is so my boss comes from statistics background so he's really good i'm not and then um i i kept looking at the data after pca before pc and bunch of stuff that are going on so i'm saying pca because now i know that the problem was pc but like a lot of steps before pca a lot of step after and i don't know there should be something going on in the pca but we cannot touch pca of python and he was like why not that's the problem with the eigenvector and was like oh i knew what i can victor is but i had no idea how to touch it and then i looked at it and then he was like you know what we don't want to prep we don't want to have we don't want you as a programmer we i i know that you're really good with python you're amazing with python but i want you to be able to find the issue that you have with your data that's your data science job and he's correct that was my main job so um yeah that was actually that was actually a flick into my head like hey that's that you don't go to your boss and like hey what is the problem with my data that's what you have to figure out so yeah that's like know the details of the algorithm that you use if you want to you can just like you because guys like because we are data scientists no one's job is waiting for us a company is going to earn money while we're doing our things and so that's a good thing because we're going to have time both for chilling and to learning so there is going to be a good time for us to go and look at what algorithm i'm trying to use what are the options i do have and uh kind of like learn what your i mean your boss definitely have some good experience and use that oh yeah and also like i think the first thing you said is it was one of the more like important points is just ask your data questions right because that's like if you don't know why you're there and what you're doing and like what the overall goal is like you're going to struggle to get their angle and also like look you obviously want to do cool data science things but to your point you have bosses they have bosses they need to make sure that you know they know what you're doing and why um chris uh chris phillips asked a question about um when connecting with people on linkedin is it better to go for just for like high level like department heads or um kind of no i go for data scientists usually i mean if the senior would answer me that's great i usually don't have time i mean i just the other day a microsoft senior answered me i mean i'm not looking for a job i just want to like learn about microsoft i've heard that they're they're not really like huge in data science so i'm just like thinking about my next job in some years but i'm trying to make myself ready for that i'm collecting data and i i wanted to talk to him and he answered me but most of the time the managers they don't have time so i go to the scientists to my level which level you whatever level you are i go to that one those people yeah and honestly like you could reach out to all of them like i mean i do for these kinds of talks and it's always surprising when i get someone who's like senior director to respond and they'll say oh this is cool i'm like always impressed but um i'd say go for a little level and then go do both i mean it costs you nothing yeah um but sorry we need to talk offline because i've got some folks i think you need to be introduced to in particular at microsoft um but also i have the stats brandon i think you you'd uh what i do is that i use i talk to data scientists because first i can ask them what will be the questions and also i can um they can they usually say like oh i know the hiring manager of this position i'm gonna go and give them your resume in person so there are some benefits if you are from that field they're gonna pay more if sorry the person who's referring you is the referrer as the data scientist they're going to put more attention into your resume yeah and if you're actually good or i mean part of it's like look you want to become friends more than like it's good to reach out but if they think it's super transactional i was gonna do it but when you ask questions and you try and be nice um ultimately everyone knows right if you guys work at companies you know that you know a lot of times people have referral bonuses i have friends who'll get like five grand if you refer someone so but also you're stating your name on this person so you know there must be something there but um i think there is also financial incentive for getting referrals so um you know it never hurts to find a win-win and then um natalia asked um how many projects do you do jeff was it just like one big project or is it like you have like 10 things on your plate i i'm sure it changed a lot too yeah i mean i don't have that much experience to basically talk for most data scientists definitely i i only had like one year and a half experience but for my experience for when i began and when i started they just put me in the small projects just to learn what's going on and not just delta but like the data science part of delta and like talking to different groups like touching different tables like changing some stuff giving me a very small project but hey for this specific origin and destination tell me what is the tell me the average of um tax that we we recharge the the customers and um then after that i had like a real defined project and it was only that and i was really bored all the time because i the thing is that with the corporation that you have to sometimes wait for very very be busy bosses to get back at you other than that you don't have anything to do and i had to wait so long for an answer and so it will take you a while and um i just i i decided to like i had that on the background but it was in a coveted situation so it wasn't a really bad thing so i was like okay i can just chill when i don't have anything to do away from my bosses but now that i'm like i'm like really into things i have two projects at the same time usually one of them has the priority and i own um the first project just by myself um and the second one is um basically the combination of different departments um but what what you own when you own your project if you have more responsibility you have to follow up you have to convince you have to do everything about it because you did you own it um and then sometimes just the maintenance of your previous project will come back to you you know you're gonna you might end up like one day with four things to four different category of things to do and someday you just have one thing to do but yeah at the same time the defined projects are two awesome well listen folks this has been really good i'm sorry we've had you for over an hour so i really appreciate it i think it's really good i think we got a lot of good uh good insights good advice um i put my linkedin uh in the comments for anyone who wants to connect um sarah obviously you guys can go on linkedin find her wow reach out to me for referrals uh reach out to me you don't need to be fancy with me please just just tell me that you need referral i'll be more than happy to help you out that's i got my job like that i'm sure that um i understand the frustration please do that email me if you have any question you need to talk to me um we can set something up to talk on the phone um and uh yeah i mean i i'm i don't i think i can find my linkedin right now and share with you all right i'll go grab it and put it in um another thing folks um we do have uh our arkham tables that are open so anyone wants to hang out and talk to each other uh it's always a great opportunity to connect with everyone else here so feel free i think the tables are up to 10 people each always a good a good group of folks to hang out afterwards those will stay open um and then look out for emails for the upcoming events um they're gonna be really good i'll kind of similar like this so if you like asking questions and keep getting that off your chest like definitely come join again uh and of course definitely please give us a like and follow on linkedin youtube and your your profile your social channel of choice uh we're on all of them um this talk is gonna be recorded so i'll send that out an additional email you'll have um our youtube channel so you can see this talk all the other ones that we've done um again thanks everyone for showing up um i know there's lots of other places you could be on a thursday night so um always always happy to see everyone uh thanks again again my name is aaron fellus um and i will see you next time can i say one thing before we go yes sorry guys um i this is my third um third talk that i had this month and all the talks i say like hey guys if you're a data scientist and if you want to get uh get a kaggle experience i'm looking for people to do it with me please join i mean reach out to do it with people i'm i'm doing it myself now but it's boring alone and i'm not learning a lot from just myself and the internet so yeah just reach out to me if you're open to do kaggle with me that's really cool um very happy you use that because honestly like usually what happens is someone waits until the very end after we're not live anymore and they're like by the way i'm hiring and i'm like oh you should have told me um but yeah so folks actually i highly encourage you to go into kaggle and do that because that's actually a really great way of learning and honestly like if you're looking for data science job really good to talk to another data scientist who can give you pointers so not hiring but delta is hiring so check that out and let me know if you need any referral awesome guys thanks see you next time

Channel: Breaking into Data by Promotable

Views: 158

Rating: 5 out of 5

Keywords:

Id: 6bMmPukYzZc

Length: 59min 57sec (3597 seconds)

Published: Wed Aug 04 2021