Art and Data Science: Her Kaggle Grandmaster Story (Andrada Olteanu) - KNN Ep. 85
Updated: Feb 22, 2022
Andrada is currently a Data Scientist at Endava, Kaggle Notebooks Grandmaster, Dev Expert at Weights & Biases and Data Science Global Ambassador at Z by HP. She has a bachelor's in Statistics and a master's in Data Science and Analytics. She combines the artsy side of data with the technical side to create fun, educative and insightful notebooks!
[00:00:00] Andrada: All your social media presence that you regard as being professional. For example, LinkedIn, Twitter, Kaggle, if you're on or any other platform, it's your kind of professional. Upwork and so on. It's good that you have the same photo all over the place. First of all. And second of all, I read this and it's actually makes it a lot more sense and that you don't change it very often.
[00:00:34] Ken: This episode of Ken's Nearest Neighbors is powered by Z by HP. HP's high compute workstation-grade line of products and solutions. Today, I had the pleasure of interviewing Andrada Olteanu. Andrada is currently a Data Scientist at Endava, Kaggle Notebooks Grandmaster, a Dev Expert at Weights & Biases and Data Science Global Ambassador at Z by HP.
She's doing a lot. She has a bachelor's in Statistics and a master's in Data Science and Analytics. She combines the artsy side of data with the technical side to create fun, educative and insightful notebooks. In this episode, we learn all about Andrada's experience growing within the Kaggle platform. We also touch on what it means to be a Z by HP Global Ambassador.
Finally, we finish up with some fun Romanian idioms that we both had a good laugh about. I had an incredible time speaking with Andrada and I think you'll enjoy our conversation. Andrada, thank you so much for coming on the Ken's Nearest Neighbors Podcast today. Obviously we've had some interactions we've we've been on panels together and it was long past due for us to do this. So I appreciate it.
[00:01:36] Andrada: Thank you so much. Thank you so much for heaving me. Thank you.
[00:01:39] Ken: No problem at all. I had to get you on, you have a really cool background, you know, a Kaggle Grandmaster, or you've worked on some interesting projects and I've loved hearing you speak in the past.
So this'll be a nice fun conversation, and I'm excited to be able to share some of your insights with the community here.
[00:01:59] Andrada: Okay. Thank you. Thank you. Thank you also for your contribution for this community. It's absolutely amazing. And we are glad to have your #66DaysOfData challenge, data science challenge, and like all the fun projects you're doing and your YouTube channel.
I remember the first time I kind of encountered you. I think I've already said, told you this was a few years ago back when I started to doing data science and I didn't know how to scrape it was a scraping code video. I don't know if you remember doing something like that, but it was...
Something like that.
And I kind of coding along with you because I had no idea how to scrape. It was the first time I heard the word scraping and it was super, super fun. And then to realize we are ambassadors together and you're also active in within the data science community is absolutely amazing. Yeah, so thank you so much.
[00:03:03] Ken: No, that means a lot to hear that, and you know, I was talking to someone the other day and it's funny, I don't say this that often, but the creation of a lot of this stuff, it's a little bit of like a selfish. Right. Like, I want to learn new things in the best way I learn is by teaching. And like the beauty of the system is that I can like selfishly get something.
I can have conversations with really interesting and cool people like yourself, but other people can also really benefit from them. And to me, that's such like a beautiful thing is I can get what I want. I can pursue the things that I'm passionate about, but there's this other like side benefit where other people get a tremendous amount or hopefully get a tremendous amount of value out of it as well.
And I like the idea that there are systems out there. We can be happy. We can optimize. We can even make some income from a lot of lessons as well, and we can also create good value. That's free, you know, or majoritively free, which is, which is such a, such a cool thing in the world today.
So I I'd love to jump a little bit more into your background. I appreciate all of the compliments, but this is about you, Andrada. So the, the first thing I always like to ask people is, so how did you first get interested in data? When did it come on your radar? You know, was it something that happened like, Oh, I saw this really cool project and I got excited and I knew it was for me. Or was it more of a slow progression?
[00:04:33] Andrada: So frankly, to be talking it wasn't something I knew I would do. So when I finished high school or my colleagues, I think 50% of them knew from the beginning of high school, they wanted to be doctors. The other 40% wanted to be engineers. And there was, there was another 10% that was already kind of knew what path they would take.
And I was completely like in the sky, I didn't know whatever I want to do with my life. And I kind of crossed out all the universities I got into statistics just by poor chance, I think, and then struggling to kind of find a job because this was something I really wanted during my high school years was to kind of get independent, make my own money and just.
Trying to provide for myself. And this was extremely important to me back then. And it's still now. I got the job within data analytics, so I was a trainee in data analytics, and I started liking that especially the part where you do data visualization and present. The business cases were also super, super interesting just because it was a fun, like activity, not getting too much into numbers, not being math related, but still being kind of very artsy.
And I would say kind of clever or I don't know, it was super interesting finding insight and so on. Yeah. And then I realized the next step would be data science and it was like these two parts before me either doing data analytics or doing data science. And I got a little bit acquainted with it.
I decided to do a course and then pursue a master's. And then I kind of, this is me now, so it's no going back, but I really, really love how everything turns out. I really do think that some things I kind of, you know, when you have something in your back in the back of your head thinking, I kind of want to do something great, or I want to do something artsy, or, you know, kind of what you want to do, but you don't necessarily know what's the job.
And I think in 2012, I don't know if data science was even a thing, this big thing as it is now, and I've never heard of data science. And we had "Statistician" kind of within job descriptions. That was the only the closest you could get or computer science. But my journey kind of carefully throughout years got me, got me here.
[00:07:44] Ken: That's awesome. So something that really cracked me up is that you said you got into statistics by chance. That was that was a funny one or the probability of you landing in statistics.
[00:07:56] Andrada: It was crossing out university. So I had a huge list of all the universities within Bucharest and I just crossed out everything and I was left with two or three things and a very good friend of mine within my class.
He was going to statistics, and I was like, Okay, let's do it, we shall see.
[00:08:22] Ken: Well, it worked out randomness worked out in your favor a little bit there, which is pretty cool. I really I'm interested in that concept of sort of knowing what you want, but not knowing how to call it or what it looks like in practice. I think that that's something a lot of people, a lot of people feel, I mean, I know I felt that before I started data science, before I started content creation, any of these things I said, oh, okay.
You know, something that always fascinated me was sports and understanding sports better. And, I found over time that the best way for me to do that was with data and a lot of this information, but there was no field called like sports understanding. Right. And I also loved teaching and understanding that, but frankly, I was like, teachers don't make any money, at least in the States and like there has to be...
So I was like, okay, well, how, you know, the way I've found to do that is making the roles for myself. Right. But, you know, how did you eventually like come to this realization that, okay. Data science is that domain where I can do that? Or like, where are these things in the back of my head come into fruition?
Or was it that, Hey, within data, I can kind of shape my role a little bit to have exactly what I want. Was there, you know, how did that connection happened.
[00:09:54] Andrada: I think it's the second. And you kind of, you touched right. What I wanted to say. The fact that I don't think it was what I wanted. And then the role kind of came up was the road.
And within the role, I kind of navigated and chose. And I think everybody does this. And this is absolutely amazing in this world nowadays, is that you, there are some jobs that were never before, like YouTuber, content creator, you can make money off Twitter just by off Twitter, just by talking. I don't know, what's the maximum 150, 100 characters or something like that within a tweet.
And you can do that, or it's amazing. It's amazing. And the fact that you could take something like data science, which is so broad, like you can, as you said, you can have teaching in there. You can have maths, you can have artistic stuff. You can have coding. It's so many things that you can choose what to shine a light more or less to make your own concoction of what you want to do.
So I think in my case, and maybe many, many other people's case is I kind of added my own, let's say, personality into this role.
[00:11:34] Ken: It's cool that there's sort of two sides to that. So you can add your own personality to the role and study and learn things that you're fascinated with. Right. And that makes it so that the work is more interesting for you.
But on the other side of that, if you're looking at it from a company perspective, if someone's specialized in something and they get really good at this one thing, that's also really valuable and beneficial to you as well as like the organization. And so I think it's nice that there's, that incentives are really aligned for data scientists to say, Hey, like I get in, in a more generalist role, I'm interested in pursuing this path because it's interesting to me, it appeals to whatever it is.
And then the company, hopefully they're, they're willing to like, let you have some freedom because it means you're going to be doing better work. I think there's sort of a magical balance there. And not all companies are like that. I'm a realist, but a lot of companies do let you have some freedom. I mean, I I've a friend over at one of the FAANG companies.
And she's within the company gone through like two or three different roles already. And I think that that's, you know, because she's like, Hey, I want to work on this. I want to work on X, Y, Z project, and they let her do those things. And to me, that's, that's really cool. Right? I mean, the companies, obviously any of the FAANG companies are big enough for, well, you know, what do they care?
Their teams are massive and jumping around. But I, you know, I would encourage that if any, like employers are listening to this, I think that there's tremendous power in that flexibility. And it's just better off for everyone. Right.
[00:13:16] Andrada: Endava, which is the company I'm working at the moment is exactly the same in a sense of and this is why I kind of love that I got the opportunity to work there just because they have so many patients.
And they work with multiple clients or they can come from banking, automobile food industries, and on, so it can be so many, so many industries, but there could also be so many projects. You can work on an NLP, you can work on object detection, you can work on video data and so on. And as a data scientist, because it's so broad, and I had, like I said, I had, I had a friend, she was obsessed with computer vision. She loved computer vision, but when she heard the no patients like, oh no, she hated this area. And I was like, but it's still data science, like Nope, images, images. So the fact that in Endava and any other company can have the ability to let their data scientists kind of move around and gain experience really from one project to another area, to another.
[00:14:32] Ken: I love that. So I'd love to hear more also about sort of your story of landing that job. And was that the first job that you had out of your master's program? What was that experience like? You know, I think that there's, that's always very interesting for a lot of people now, you know, is Endava national companies, a local company.
[00:14:50] Andrada: It's an international company. It's based in UK, it started in a UK, but now we has offices all around the world. And it's kind of expanding. I think I kind of owe all my little and bigger jobs to Kaggle and this is why I usually repeat myself or all over again, just please, whatever you do, if you're a data scientist, take advantage of this platform.
And in any really career, if you're a designer, just go over to these platforms, open sharing platforms and share your work and receive like attention and give attention and learn from other creators and so on. Or there are so many platforms for so many, so many jobs and Kaggle for data science, proficiency say it's not really a job as a profession.
Kaggle is absolutely amazing. And I started can go by mistake really. The course I followed within data science at the end. They were like, Oh, you maybe want to try cake. And I was like, Yeah, let's do this. I kind of did all their courses. I found out what the notebook is, kernel back then. And I just started working on a project just because I wanted to learn more.
And then. Stuff starting started happening. I found out about medals by mistake also because, so I started my master's and the first module was crazy and I was just getting adjusted. So I haven't done turned on Kaggle in like three weeks, let's say, and I had the notebook on Brazil fires, which wasn't even that good, actually, like in my head, I was just having fun with seaboard and I was like, Ooh, what's this?
Okay. Let's see birds. And I went back to Kaggle. Once that module was over and I had the gold medal, I was like, what is this? And then I, I realized people actually unfold. Wait, where did these people, how, how did they find this? I think the data said became. And then kind of the notebook started to go up.
I think this was, this was, this was the case, but then I realized I can actually interact with people. And I found out there is a community on Kaggle and then I got into competitions and so on. And so my activity and kind of going up and starting to kind of navigate and interact with the platform and understanding what I can do and what I can't do or what they shouldn't do.
Got me more opt into the notebooks field and then that'd be kind of found me. I was a fit and girling over. I haven't fixed it to him. Although he was also, he's also a Romanian. I was like standing in the back, his straight times Grandmaster. Now I can't, I can't even look at him, but I'm just going to follow him and learn from him.
And then he kind of find out one of my notebooks. And when I when I finished my masters I contacted him and asking me he had a position open and he had, and I had a few interviews, talks with HR. I didn't kind of get the right thing to it. I didn't receive any special attention, unfortunately.
But yeah, I don't think if Kaggle wasn't on my CV or resume, I think you call it I, I'm not very sure I found out even about Gonzaba. Yeah.
[00:19:13] Ken: That's awesome. You know, something that I don't think people realize is how powerful algorithms. So you think about it on Kaggle, right? And maybe it's not a machine learning algorithm.
They use, it's a popularity vote, whatever it might be, but there's momentum that you get from platforms sharing your work rather than you just sharing it. Right. So if, if I make a website and I don't optimize it and I put it on the internet, probably no one's going to watch it. But if I make a tweet or I post something on Instagram, I post something, YouTube, I put something on Kaggle.
If it's good content, there's incentive for the platform to share it and perpetuate it beyond just me. Right? Like, you know, I make a YouTube video and without having to do anything, the platform shares that thousands of people, right. To me, that's fascinating. And that's a great way to create scale in your work.
Kaggle works the same. If you've produced interesting stuff, you're working on a interesting dataset. You have a halfway decent analysis, the odds are, someone's going to look at it, right? And then someone's going to find it. And that discoverability to me is like a huge, important part of, of the job search.
So rather than going out and doing all the hard work of finding things, creating systems where you're findable or more things are coming to you is something that I put a huge, huge premium on, or even not like being the first thing that comes up. But being searchable is really important. I mean, I don't know a single employer who doesn't like when I was interviewing candidates, I would Google every single one.
And it's like a very easy thing to do when you get a feel for what they're about. You see any interesting news headlines, whatever it might be. And I, you know, I love the Kaggle platform. I think that that's such like a, an awesome way to start. What were some of the sort of challenges with Kaggle that you faced starting out?
You know, obviously, even, even now, I mean, I have a couple of Kaggle notebooks out there, but it's still intimidating for me to enter competitions still intimidating for me to, to commit that much time to it. You have any tips or tricks.
[00:21:28] Andrada: So having, like interacting with cattle and having, let's say a presence on there, it's extremely important.
Not only because you are learning and if you, if you really want to learn more, that's the best place you can do that. But also because you kind of have official kind of a portal. Because I can, for example, say on my resumes, they know how to inspire touch for example. But if somebody goes on the Kaggle platform, searches my profile and then finds even 1, 2, 3, 5 notebooks that use PyTorch, they kind of feel eyes my level, they foresee, Oh, she knows a little bit of PyTorch.
First of all, second of all. Oh, she has a certain level. Maybe she's beginner, she's intermediate, but you kind of say all this information within a resume, so have a feel of the person and also the commitment also on, so it's, it's a great bonus. Some things that I kind of did wrong when I started Kaggle.
So the first thing was, I didn't even the first year, I think, or half of the first year I was committing blank. And why did I do that? I didn't even know what the commit was. I haven't worked with gates yet. And I was like super, super confused for me, committing was saving a cherry and that was, and I saw some other people.
So this was early, early beginning for me. I saw some other people kind of committing without changing their notebooks and they are kind of going up above me. And I was like, but why is this working? Why I think the Kenalog kind of sees that you are doing changes. I don't know if this is still a thing on Kaggle.
I don't know. I am not very sure. But back then, if you would commit a notebook daily kind of, so you had activity on them and they would keep your notebook up more, let's say because they were like, Oh, she's actually working, whereas this other person, they worked on this notebook today. Okay. That's some big no-no, it's very, very, very weird and it's extremely the sound productive.
And when I realized this isn't really a very good thing to do, I kind of stopped. So if you see other people, because it's kind of a feeling when you see 100 people next to you doing something you're kind of incentivized to, okay, I need to join the group, but usually that's not the case. You want to do the opposite of what the kind of every bigger group is doing or something like that.
The second thing I had to realize is not a very big, it's not a, it's not a bonus was kind of. This was also early days, but I saw lots of people kind of saying congratulations. And I was like, oh, I need to be productive. And I need to be kind of proactive, supportive. So I would comment as well. And then I realized that's not good either.
Actually on Kaggle, I think there's a post with a lots of tricks that you would think you, somebody entering on the Kaggle platform knows that because Kaggle is such a big platform in super complex, in my opinion, especially for somebody with like very fresh eyes. It's very good to look over.
I can't, I can't remember. Maybe I going to send it to you. That'd be great. Yeah. Yeah. And another thing that I think I did once, but then I kind of realized it's it's not okay. Because I saw others do is extremely important to just reference, reference anything that you, you use, especially if you're using another notebook.
It's extremely, extremely important because you're kind of transparent that you didn't invent the wheel. Other people had to, and it's healthy to kind of show that you are also learning from other people. Nobody's God on Kaggle, everybody's learning from other people. I feel like it's always a kind of, if you watch you read one of my notebooks and there's a piece of code there from somewhere and that that's somewhere.
Connected to maybe other two sources and then it becomes a tree or kind of knowledge. So it's extremely important to reference. Yeah. And vote, if you fork a notebook, this would be another, another very, I sometimes forget to upvote and I realized, and I go back and I'll put with it. It's extremely annoying.
[00:26:45] Ken: Yeah. Well, you know, the accreditation, I think is something that is, is super overlooked and I've, I've, I've thought about this from a, like a, like a practitioner perspective or someone who has a lot of public facing work on the internet. I really see zero benefits of not sharing someone's work.
Right. If someone, for example, a couple of people made reaction videos, right. Or they talk on the same concept and you know, there's 20 videos out there. That's like how I learn XYZ. If I started it over again, if someone ever references me, I just share the video to my audience, which is probably bigger than theirs.
Right. It's a no brainer for me. It's like, well, you know, there's a, an authority aspect where someone is, is, you know, viewing my content or viewing whatever I've published. That makes me frankly look good because they're like, Oh, this is interesting enough that I'm going to use it or respond to it or whatever it is.
Right. And there's also the aspect of just like the good view that it creates. You're linking yourself to someone who's probably a lot more accomplished than you are, whatever it is. And so to me, I've never understood why in this space, you would try to claim someone else's work as your own because it's so easy.
It's so easy to figure out when something's not yours, because people will call you out on the internet, you know, like, well, what's the point? It takes five extra seconds. You already have the other page open in your monitor, just like copy and paste the link at least. Right. Like
[00:28:28] Andrada: copy and paste the link.
Yeah. And and you just I just remember the first time I properly referenced somebody was when I learned how to do sentiment analysis on Rick and Morty's tweets. So, Rick and Morty, I don't know if it's since the series, but yeah, I kind of found that the tweets script, sorry, I kind of found the scripts from some of the seasons and I kind of do sentiment analysis and I referenced Savier. I don't remember if it was from Spain or Portugal, he did this amazing notebook for I think it was Lord of the rings and I basically copied the entire notebook. And then I just added my own field, some more stuff, what, it was very feel very similar to what she has already done, but on other data sets, but this is what, this was my purpose.
I was astonished, but by what he did, I was like, I wanted to learn this. I would be very grateful and I'm super grateful that he shared everything. You already did all the research. I don't have to do it. So I'm going to start learning. So I kind of did a code. Line by line. I didn't copy anything. I was coding alone along with peace notebook.
And when I was ready to share it, I already had two back then and I found him on Twitter and I texted him and I said, here is a notebook. I am super proud of. I loved working on it. All the prices should go to Savier because if it was, it would do nothing his notebook, this wouldn't have happened. And I remember I didn't expect it. This, he, they shared commented and also DMed me within Twitter. And he said, Oh my God. You made my day. I'm so grateful that I could help you because he was excited to sweat right here is this girl that she's a noob.
She doesn't know anything, and now she knows some stuff and it's because of me, like why would he not be happy? And I kind of then linked a little bit of friendship because whenever I would need his advice or a problem, I would just message him a very quick question, and he would answer. It was absolutely amazing.
And if I wouldn't have referenced him, I wouldn't have had this kind of interaction. Yeah.
[00:31:34] Ken: There are so many really powerful things in what you just said there. So the first, I think is important for people to realize about just projects in general, like a lot of projects are taking something that exists, crediting the person who created it and taking it a step further.
Right. You know, let's say I did an analysis on the Titanic dataset. Right. I have a full tutorial on that on my channel could probably be better. I've learned some stuff since then, but at the same time, if someone takes that, you know, they, they take everything that I did and they add to that analysis.
Maybe they try some different models, maybe they do what you know, maybe they create some other visualizations. Maybe they, they expand on it. Like that is a perfectly good project that you can share. And as long as you're crediting the source material, same thing with just applying an analysis that you've seen to other data, like someone like I did this analysis.
You know, I probably shouldn't say this publicly, but I scraped a bunch of Glassdoor data on data science jobs. Right. And then I analyze that information to see, okay, what are people looking for? What are the skills? What a lot of these things, if someone did something similar, but they used maybe not data science jobs, right.
They, they looked at consulting jobs or they didn't use glass. They used another website, but essentially all the methodology would be very similar. That's a perfectly good project in its own. Right. And again, if you're like crediting the office, It doesn't have to be this completely unique, original thing.
You can either change the data that you're using, or you can add to the analysis and, you know, use different models, whatever it might be. And that's what makes it unique. So many people get so overwhelmed, they just don't realize they can start with a foundation and expand into it. And then, you know, if you do credits it so on and they can create these great friendships that, that perpetuate even further.
[00:33:30] Andrada: I would also like to add to that. So you said people are like scared or and I think I'll have to talk at some point and was also there and she said something that's quite, I didn't really empathize with it, but I kind of saw that this was actually, I didn't realize this was a problem. So what she said was that she would receive very frequently. I think I'm not butchering what she said. I hope she said she would receive many messages from girls that would kind of give her their notebook before sharing it. And they were, would, they would do that just because they were very, very scared or under pressure being like, Oh my God, I'm posting this notebook.
What will everybody think? Maybe I did a mistake and so on. And so they are, they were kind of trying to get an approval from Peru saying, okay, you're, you're fine. But she, she, she actually encouraged them to just post it because nobody is going to be mean with you if they are there's, there's their problem.
But I haven't seen really on Kaggle police. Like I really, really actually know there and I'm thinking no bullies there. But is this kind of pressure that you need to be very unique and outstanding and revolutionary even that's they feel maybe like this is not original work. I just did very, very little to add to this.
And why would I deserve to post it? Which is not the case. You are first doing this notebooks for your learning. So it doesn't matter, even if you didn't change anything. So maybe I just copied your notebook from 0 9 0 2 1 line 100, like line by line. I didn't change anything yet, but I kind of realized what everything is doing.
And I learned maybe on your technique and so on, you don't really have to change anything. You are learning and coding along with other things. And then if you want to take a step further, you can kind of add, like I said, your sparks sweet, but you just shouldn't be, you shouldn't be afraid. And I see why you could be, but you shouldn't really, it's, it's a very nice forum.
And it's an opportunity for other people to kind of jump in and say, Hey, maybe you get something wrong here. And then we can say yes or no. Or why did you think that the purchase on it's a conversation?
[00:36:19] Ken: This episode of Ken's Nearest Neighbors is brought to you by Z by HP. HP's high compute workstation-grade line of products and solutions. Z is specifically made for high-performance data science solutions.
And I personally use the ZBook Studio and the Z4 Workstation. I really love that the Z line who comes standard with Linux and they also can be configured with the data science software stack. With the software stack, you can get right into the work of doing data science on day 1 without the overhead of having to completely reconfigure your new.
Now, back to our show. I love the idea of feedback. That's how we get better. Right. And to have a place where you can get structured feedback where someone's like, Hey, why did you do it this way? I get that a lot. And I actually have to think about it, right? Some of the analysis I haven't done like two years ago that I did it.
Right. And I'm like, oh, hold on. Let me look into the code. And then other times, you know it's just nice to have something out there. I mean, you were searchable or whatever. It might be something that I've realized is the more that you post on a platform, the less you worry about those types of things.
Right? So your body of work should speak more than an individual thing that you've done, or you've done quite a few Kaggle notebooks so far. I think each individual one probably means a little. Less to you in terms of like your intimidation about sharing it or because it's been out there for awhile.
Right. And so to me, that's something I think is really powerful. So a couple of days ago, someone reached out to me, I put a video out and apparently there's like a typo in one of the screens in my video. Right. And, you know, they reach out and I'm grateful. I like when people do this and they said, you know, like, I wanted to let you know before, you know, you know, shit hits the fan, you know, you know, I don't want you to get criticized for this.
And I'm like, what? No one is, if I have a type one, my video, no, one's going to say anything. Right. Like just let it ride. It's fine. You know, if people notice it great. But it's a small thing who cares. Right? I mean, I didn't phrase it like that. I actually am very grateful when people, because if it was a big thing, I like use terminology wrong or something like that.
That'd be worthy of taking it down. But you know, there, there are certain things that you're okay with once, once they're out there. And once you started sharing with like, you know what people are going to criticize you for, right. If I did math wrong on a video, That would be something people would, would roast before it, and not like one plus one equals three math.
Like if I described a concept wrong, if I said, oh, you know, like linear regression, like multicollinearity doesn't matter or something like that, I would get roasted. Right. Cause that's like a professional domain, but, but you know, like those are things that you learn over time where the threshold is. And the only way you learn is by putting stuff out there and experimenting with it and, and it should be fun, right?
You're like engaging with a new community and you don't have to like talk with them. If you're an introvert, they're just, you know, people behind computers on the internet, it's just, it's, it's a Twitter, right. It's fun. You engage, you share. And there's a chance to go viral. Kaggle a lot more friendly than Twitter, I'll say, well, you know, there is some power in Kaggle.
And I also think on LinkedIn is there isn't as much anonymity on those platforms. Right?
[00:39:44] Andrada: Oh, what I wanted to actually share, like from good practices. I don't know if it's only Kaggle, but all the social media, I don't know if people know these, maybe they know this. It's all your social media presence that you would regard as being professional.
For example, LinkedIn, Twitter, Kaggle, if you're on or any other platform, it's your kind of professional. Upwork and so on. It's good that you have the same photo all over the place. First of all. And second of all, I read this and it's actually makes it a lot more sense and that you don't change it very often because people don't use journey like read your name.
They look at the photo and I realized, because I was doing something like that as well. And at some point, boy , am I pronouncing correctly? Hopefully.
[00:40:54] Ken: Better than mine.
[00:40:54] Andrada: So yeah. He changed his photo for a day and I was like, who's this guy? Ah, he's okay. I don't want don't follow these guy.
But he was like, oh, so weird. And then she kind of changed back and tweeted about this breathing that it's good to leave your photo alone because, or if you change, it just don't change it once a week. Change, try to change it once a year, maybe two years and so on because people kind of look at, for example, tomorrow, if I take your headshot from two.
And I replace it with mine. I might get followers just because people want, raise my name, going to ah, Ken. Oh, I don't, I don't have any follow up follow. Yeah. So please don't do that. Okay.
[00:41:46] Ken: No Ken in person. On YouTube, there's a problem with a bot impersonators. Cause you don't, you can like essentially make very similar names.
And so it hasn't happened to me too much, but there are a couple of my friends, there are bots that copy their, their name and will comment like crypto scams in the comments, which is, is nobody. I think YouTube is cracked down on it a little bit, which is a good thing, but it was a a problem for awhile.
So, you know, kind of continuing down the, the cargo thread. So how, so you're a Grandmaster in notebooks, correct? How did, how did that happen? You know, like what is the progress for that.
[00:42:29] Andrada: So as I said, at the beginning, it was just making no books and kind of understanding some techniques, data, even if it was just something interesting, I would do it.
Then I kind of realized, oh, there's metals. And I memorized, I think how many times I ended up into the scheme from Kaggle, when you see from no books, competitions, data sets, and so on how to get to different slab, different levels, all the way to Grandmaster. And I was like, oh, okay. So I get, have to get to master.
I think it was 10 simper medals. And so at the beginning, as I said, it wasn't about metals. And I was like, Oh, okay. So maybe I can become a mustard. This sounds interesting. At the beginning, when I started, when I entered the platform, I was like, I don't have a chance. Like I I'm a novice, I have four no three, like circles.
It's like, how do these people do this? But I think, I think we are in an era where we are so used to making things very, very fast. Don't climb the stairs. We have an elevator we have cars, we don't walk. And so on. We have instant messaging. We don't have to kind of write letters and so on. And I was like, Oh, I have to do so much work.
And it's, it's true. You have to do a bit of work in order to get to two grand master level. But I was like, oh, I don't have any chance because I was like very, very, very used to this very quick thing. But then when I realized that's not impossible, I kind of tried to expand my work and every notebook. I tried to make it extremely different from the other one, which actually required me to learn something new or at least
very new or new within each notebook. Whether it was, I had one notebook, which was Mastering ggplot and all the things you can do in ggplot, which is R. I kind of dabbled between R and Python. It was sentiment analysis, which was another thing. It was computer vision. My first stake into computer vision with melanoma, which is again, extremely new.
It was the bird call competition, which was audio files. Was that, oh, okay. Oh, this is so interesting. It was completely new fraud detection. So I tried to kind of think point all these beats, trying to get these notebooks. Now, some of them wasn't that great, meaning that, although I was super, super excited about them, I didn't get the attention that I thought I would.
And other notebooks just skyrocketed. I think the thing that people should really pay attention to as I said, besides the fact that you want to do a quality work. So this is, this is a must. You want to do a quality work and you want to kind of put your effort into it. But afterwards you want to take advantage of the community and sharing.
And I shared on Twitter, I had 50 followers for four months, I think. And I was sharing with my 50 followers. Some of them, they weren't even within data science industry. I was like, I hope you're excited as I am about my new units. I didn't know it was doing anything. But people are going to kind of find you and see that you are into data science and kind of start to you are starting to build a community.
You are going to start to build a community. So take advantage, do a very, very good job of creating the notebook. And I think, I think that's would be the recipe or at least my recipe on how I became a Grandmaster.
[00:46:59] Ken: That's awesome. I think that there's a lot of, a lot of connection between the different social platforms.
I mean, I kind of have if I need to start actually doing more notebooks because if I make a video on it, I have a shortcut cheating path to any with with my audience. You know, if I have I'll make a video on it, there's, you know, people are more likely to go to it and drive traffic and there's some sort of optimization equation associated with that.
But I think for, for anyone, right, it's about just dedicated work, doing things that are interesting. And also I think that there's power in finding sort of a niche. I mean, a lot of the work that I do that people would find interesting is going to be more for beginners, right? If I'm creating content on Kaggle where there's a companion video that goes along with it, frankly, there aren't that many people that do that, especially at like the big interface.
So I'm doing something that's fundamentally different than other people are doing. And that's going to be something that garners upload, right? Like if it's, Hey, this is, this person went through, not only did they make the kernel or the notebook, they also made this companion video, you know, there's, there's incentive to make something like that rise to the top. In your case, I know from looking at a lot of your notebooks, you really enjoy the visualization part. Right? I, if I open a notebook of yours, I could expect like, oh, there's going to be something like visually appealing that I haven't seen before. I remember going through, I think you did a sentiment analysis or something of the sort on Elon Musk tweets.
And if they impacted the price of Bitcoin, right. Or doge coin, I can't remember exactly what it was. But to me, I felt like I got such a chuckle out of that notebook. I really enjoyed it. You you've mentioned it a couple of times, like the, the art associated with data science. Can you talk me through how that, that kind of shows in your work and what that, what that means to you on, in the Kaggle sphere?
[00:49:09] Andrada: So. The arts came from two parts, I think. So it was the first part, just my love of painting and taking pictures and just seeing beautiful things. I just love that in particular. So that designing part of me kind of made its way to through the, the science. But then the second part was the fact that I was reading no books and although there were extremely insightful and I learned a lot, a lot from them.
I still struggle with this to this day. I couldn't understand what they were doing just by reading. And I feel like as a tech technical person, it's extremely important for you to be able to read somebody else's code. Now I can argue that I am very little experienced. This is why I can't really read the code, but this made me kind of try to explain all the little bits that I couldn't, I wasn't actually able to understand.
So this part is also in my opinion, extremely important that I'm trying to, to get to this in every notebook, trying to comment everything, trying to explain everything. If it's a big chunk of code I might be doing, I might be doing a schema just because next day, when I go back and I look at 50 lines of code in a function, I can't remember what they did, but the skim, I was like, oh, this, this, this, oh, okay.
So I already know. So it is for me, firstly, and then if it helps somebody else, absolutely amazing. So commenting, clean code, clean variables, just understanding. It's somebody else is going to read whatever you wrote maybe, and we'll try to follow. And if it's not, if it's a beginner, it's going to be very annoying.
Now I understand if it's an advanced person, they are going to get very bored of all the explanations and all that. I also like to print whenever I do a function, if it's a little bit complicated, I like to do, off-topic a very literal example of how that function works. And I give an example of the first variable and the second variable, for example, and try to sprint and say, okay, this is happening here from this.
It goes to here or something like that. And I feel like it's extremely important for somebody that can to read and understand in one go somebody else's coding. So this would be art scene in my opinion, just because it's a way of how to structure the. Actually wants to read this year book on how to code beautifully and properly in order for somebody else to be very, very easy to be followed.
Yeah. And the second part with data visualization, I just, as I said, if I like painting, and this is my, my, my faking parts with the notebooks.
[00:52:36] Ken: Well, you know, it's, I think it's one of the most important skills you can have in this domain is to write code that other people can read. I mean, you have code review, right?
Like that's something that, I mean, you're probably going to get promoted. If people can read your code and every time they're murdering something, it makes sense. Right? Like, to me, that's, there's a huge value, huge premium on that. If you're always writing code with the intention of other people, seeing it, or that's the feeling of.
I think that that's how you write high quality code. I know for my work, sometimes I know I'm the only person who's going to view the code and it is so bad. I just like do it to get it done. It's not, I don't comment it. Well, I don't. And then I, I'm the one who looks back at it later and I'm like, oh man, why did I do that?
I have to like reinterpret all of this. And it's like, okay, like the trade off between the time they're saving now, and the time you're saving your future self or someone else, I mean, that might be an interesting way to look at it is I'm always writing code for my future self. And so I should be kind to future can, would be, that's a kind of a funny way to look at it.
That would probably be effective. I, I would like to think, I try to do a good job of taking care of my future self in terms of like my. Health and nutrition and those types of things. So why not with my, not with my coat?
[00:53:54] Andrada: I don't remember. I remember I had a, so I did some coding, some functions for work, and then another colleague she came and she was like, she was trying to kind of work upon what I already did is kind of improve it.
She was like, what did you do here? And I was like ah, so embarrassing. It's so embarrassing. You can't remember, but it was months ago. So how can I possibly, but I usually at least a proper comment within that. Some functions being like this function does this for me to don't remember.
[00:54:38] Ken: And in theory, right, that is like a, a proper coding practice for each function where you use the, Oh my God, I forgot what it's called.
It's the triple quotes. Yeah. And within that, within the home, yeah, there has to be some special names, but but I mean, that's something I like literally never do and it's something I should do. Right. Look, whenever you do find a function and you're supposed to like add metadata for what it does, but who knows, you know, it's a work in progress new year, new year, new me.
So something else I wanted to touch on. I mean, we're both Z by HP Global Ambassadors and I think a lot of people don't really know what that means or what it entails. And you know I even have trouble articulating it. So I figured I would, I would leave that, leave that responsibility up to you. I'd love to hear more about that, your experience with that, and what that means to you.
[00:55:37] Andrada: I think this is by far the most frequent question I get on LinkedIn and I usually I don't know how to answer either. I think I being a casual as Z by HP ambassador it's it's first comes with such a great opportunity to work with people like you, for example, for people that are
people that are Grandmasters within multiple areas we have, we have so can you, like, we have people that are super, super experts in competitions. So we have people that are super experts in datasets, in no books, in just creating content and just growing the community. So we are a very, very interesting cohorts team.
Let's say, let's say team that kind of adds volume one to another, through the different areas. There would be experts at. And we get to do a bunch of cool stuff. Like this thing out tools like software tools, this thing out gear hardware going to events, digital events, maybe at some point, corona is going to leave us alone and we can go to actually events.
It would be so, so, so awesome to finally, like when I started my data science journey, so it was September, 2019 and I had three or four months in order to kind of start. And then this pandemic hit and I feel like all my data science career has been online, like before data analytics. Yeah, it was, I know people like physically, I but within the science, I feel everything was so digital.
That would be awesome to finally meet you with some of the in person, but we get to go to some events and get to talk to interact and everything we do. I feel it goes back to the community. So I wouldn't say we do something just for ourselves. We do something to create a better community and grow and help other people with better tools and better like techniques and gear in order for them to do what they love, which is.
[00:58:21] Ken: Yeah. I absolutely love that. I think for me, something that was really special is, you know, Z by HP, it seems like they're trying to consolidate a lot of people that know a lot in this space and get us to share information as effectively with each other and with the communities as possible. And to me, I think that that's a really cool initiative.
They're also enabling us, as you mentioned with, with some pretty cool technology and, you know, I don't look at myself or, or any of us as like like super special, right? Like anyone can spend a lot of time on Kaggle, can have a presence and, and be as in part of a cohort or a program like this. Right.
That, that idea is, is fun and inspiring. And I I'm grateful that I could have an opportunity like this as well. And I also think it's a really interesting and solid move on on HP's part because. They've been really receptive with feedback that I've had. I mean, there are times that I've been critical of decisions or hardware or whatever it is.
And they listened to me, which I think is really cool. And, you know, they they've been using us for feedback and those types of things to make better products for data scientists. And, you know, I don't know what other companies are doing, but to me that was something that I'd be excited about working in, right.
Like one of my goals is to make data science, like a better field, right. Is to make it more transparent is to have more standardization around it so people can understand it better. Cause it's a little bit disjointed as we talk about like different definitions and those types of things. And a big part of that conversation is like, it's hardware is, is like, oh, what technology should I use?
Or, or what is this? And what is that? And it's cool to be part of those conversations as well as, you know, I have only so much. Power myself to make decisions or to describe things. Other companies have a lot more big companies, you know, public companies have power to, to shape this in ways that I never could.
And so partnering with them, working with them in my mind, gives me more of a voice, more of a platform to hopefully influence the field for the better, which again is really, it's kind of special to me. I'm again, grateful, super grateful for these opportunities. And to that point of giving back to the community I've given away to ZBook Studios, you know, $3,000, $4,000 laptops to the community, giving away another one in a couple of weeks, whenever I figured out how to ship it to India.
And so yeah, that, that to me is awesome. Like they're invested in, you know, making people or giving people better access to data science, which is, which is so incorrect.
[01:01:14] Andrada: It's actually amazing what you said, the fact that to kind of giving, giving the little man power to do something or to impact something bigger.
Frankly, next to a very, very extremely big company, we, as you said, we might not have the kind of impact that they would, but they are indeed so supportive and like, they had been fun. Like this is something I didn't expect it. I didn't, and they are so fun. And so kind of like, like you said, they, they actually liked critical feedback because this is what helps you grow.
[01:01:54] Ken: Yep. So, yeah, I mean, again, I find it a fun opportunity. I'm probably going to talk to them about. You know how to make the application process for those types of things, more public and accessible. I think that that's something like if people know that this is an option and you know, they have certain credit qualifications on Kaggle or on some of these other platforms I think it's an important, and it could be a good incentive, right?
Like I look at this in a very small way of being like a sort of sponsored professional athlete. And that's so cool. Like in other domains you can, you can get that right. Like Twitch streamers have sponsors or, or they have people partners or whatever it is. Like, why not, why not data scientists? Like, isn't that cool.
Isn't that exciting that maybe that's an, an possibility within a technical domain. Like, I love that.
[01:02:49] Andrada: I love that too. It's like, yeah. You think only like movie stars and like set professional athletes and singers would, would, would have something like that, but you don't have to, you can, you don't have to sync in order to have these kind of possibilities, like yeah.
[01:03:11] Ken: Incredible stuff. So, so something I wanted to end on, we talked about this a little bit offline and you know, you'd brought up a couple, there were some Romanian expressions that have funny translations. So the first time we spoke, you told me happy birthday. And that was because it was the new year. It was actually in fact my birthday.
But so in Romania, apparently you say happy birthday at the new year, that's what the translation is. Are there other, again, like very funny chance that we have that conversation on my birthday or are there some other funny ones that, that you know, that, you know, being bilingual or possibly even multi-lingual they give you a little chuckle?
[01:03:58] Andrada: Actually, so, because I was, I grew up with this language. It doesn't smell funny, but when I went into my masters in England we actually tried to translate to a British person, some of our songs, and it was hilarious. And this is when I think we laugh for one hour trying to translate line by line songs.
And we were three Romanians and one, a British guy, and he laughed, but he didn't really understand because we it's funny because we know the both kind of both ways, but he kind of had had a blast too. So I, before this, I kind of asked some of my friends, if they have funny, like funny things that we usually say, and the first thing that's actually extremely common, for example, in English, whenever you see an older woman that you are wanting to pay respect to, for example, a teacher, we'd say even to our mothers sometimes some of us, but a granny or something like that.
And you salute, you say kind of how do you politely say hello or hi? Like, do you have
[01:05:22] Ken: No pretty much totally like, hello. Hello is more greetings. Yeah. Yeah. Salutations. Yeah.
[01:05:29] Andrada: We say kiss your hand. So this was the first thing we say is your hand and it's it's funny. We also say, I don't know at weddings, we don't say congratulations.
We say have a rock house or I don't know, like have a very strong house. You don't say congratulations.
[01:05:57] Ken: I'm going to say, have a rock house to my, my friend is getting married and like a couple of ones. That'll be a fun one.
[01:06:04] Andrada: Yeah. Yes. I have a rock house. If you're annoying me, I say you're taking me out of my watermelons.
This is another.
I don't know. I say... I actually wrote them down because it's extremely hard to translate them. We say when it's easy, something very easy, we say flour at your ear and stuff like that. So yeah, we have, we have lots of, yeah. We also say we, you know, I was actually talking with some of my friends from India and apparently they have the same parents as we do because Romanian parents sound very, very similar to how the average Indian parent is, meaning that wherever you ask something, for example, when you are in school and you ask something, they ask you back, oh, what you don't.
You don't, why are you going to school or something like that? You don't find this at school. And then you say, oh, I have a very big grades. And they say Oh, but this other friend, was he that he has a bit bigger grades or not? And if you say yes, they say something like, Oh, don't learn for me or something like that.
And they're very, very funny. Yeah.
[01:07:35] Ken: That is incredible. I think that's a really good way to end on a kind of pretty comical note. Andrada, thank you so much for coming into the show. I always enjoy speaking to you and I think everyone's gonna get a good kick out of this podcast and hopefully learn a lot.
[01:07:51] Andrada: Thank you so much for having me. Thank you.