My Data Science Journey (Zero to Freelance)

Key Takeaways

i wrote my first line of python back in 2013 almost 10 years ago and i sucked today i suck a little less and i want to take you guys through my journey and show you how i went from absolute beginner to now freelance data scientist if you're an aspiring data professional then this video is for you throughout this video i will go over my education courses and major milestones to hopefully inspire you and give you a realistic view of what it's like to become a data scientist make sure to stick around till the end where i will share my top 10 tips for learning data science welcome to the channel my name is dave and i make videos to help you become better at working with data let's get into it throughout this video i'll be using this timeline to go over all the items i want to talk about it includes all my education courses and projects from high school all the way to what i'm currently doing now brief disclaimer this is not a video on how to best learn data science this is just my journey now i won't cover everything that you see here because that would be a five hour video i'll try to keep it brief and focus on the things that have helped me the most and as i said make sure to stick around till the end because that's where i will give a conclusion and also give you some recommendations that i wish i knew when i started my data science journey okay so let's start with high school so from 2008 till 2013 i was in high school here you can see all the courses that i was taking behind or underneath the the purple header and basically what i learned here was all the fundamentals this is not that interesting but there are some things that actually help me with my data science career and that was math physics and english math because i had calculus which is one of the fundamentals for understanding the methods and models behind data science then also physics which basically learned me how to work with formulas and mathematical notations which is also very helpful for understanding models reading research papers etc and then of course english so i'm from the netherlands so i speak dutch and english is my second language so learning english in high school was very helpful to my data science career because as you probably know are all the courses all the content is in english so if you want to learn data science it's very helpful to be able to read and understand english but to be honest i learned most of my english while playing video games and that's why i have runescape listed up here because all the way through high school basically this was my main focus everyone used to play it back in the day at least at my high school i really liked it awesome game that's where i learned most of my english to be honest okay so after finishing high school i basically still had no idea what i wanted to do in life i didn't know what to do i didn't know what computer science was i didn't know what machine learning was but i looked at the website of the two closest universities which are both in amsterdam and i basically looked through all the degrees and just read the description and i was like is this something i would like i don't know this says something with computers and tech and stuff like that i think this is cool and that's basically how i pick my bachelor degree so i ended up picking the degree lifestyle informatics at the frey university in amsterdam and the degree lifestyle informatics it's kind of a confusing name i would say so they used to call it artificial intelligence and then they changed up the program a bit to be more focused on basically helping humans to live a better life through or with the help of technology and the funny thing is i just mentioned that i played a lot of runescape during high school and after finishing high school during the summer break i was kind of getting bored of runescape and it was summer break so i had a lot of spare time and i was used to playing video games so i thought let's look for something new and my brother had been playing league of legends for quite some time and i thought let's try it out so i just got rid of playing runescape and then i basically got hooked to playing league of legends so yeah off to a great start for university and i would definitely say first two years of my bachelor league of legends was the main focus and courses were second i got to platinum 4. so what was i doing during my first year and as i mentioned my degree wasn't all about data science so a lot of things here are not directly relevant for data science but this was the year i got an introduction to python so i wrote my first line of python code within this year and i also had a course called web technology which was also quite cool we learned html css build a website stuff like that but i remember this discourse and it it was hard i also filled it the first time so i had to retake it the assignments were hard for me and i think that's also mainly because i had no idea what i wanted to do with biden and and what it was for i thought it was quite cool but i didn't have like the intrinsic motivation i would say to go out and go beyond the stuff that was covered in the lectures to learn more about biden for me this was just about passing the course and i remember that the assignments so these were individual assignments but everyone was doing the same assignments and i was i would just ask around within my class like how do you do this how do you do that and i would copy certain pieces of code and i know for the final assignment we had to build a snake game so so the game where you go with the snake and then catch apples and stuff you get longer we had to program that and i didn't understand any of it man it was it was brutal for me but yeah managed to pass it um was my first introduction to python very cool but then i went on to the second year and this was the year where i got a bit more familiar with working with data and that was mainly through these two courses over here so databases where we got a sql so one of the fundamental skills of being a data scientist is knowing how to use sql to query and push data and empirical methods which was basically all about statistics z-scores distributions stuff like that and also assignments uh with the r programming language and again this was hard for me same as as python now we get to learn r and i wasn't really interested in it and i found it very difficult i also filled this course the first time and now now that i think about it databases as well i also filled it the first time so i had to retake both these these courses other courses you can see uh we had a lot of psychology as well academic english and also as i said the web technology behavior and health those were all not so much tech courses i would say more focused on the human side because as i said lifestyle informatics bachelor was focused on helping humans to live a better life through technology so technology was a part of it so i would say the main takeaway from this is that if you're currently learning data science and you're thinking man statistics programming i don't know i find it very hard i don't know if this is for me if you really want it and you really like to pursue a career within data i would say anything is possible it just takes some time to get used to it because it's so different from our everyday life and other typical courses that we get through high school high school and stuff at least it was for me okay and then on to the third year so my final year of my bachelor degree and now this degree was a three-year program but as you can see here it took me four years to complete it mainly because of the courses that i had to retake that i just mentioned and also because i was playing league of legends all the time so basically any class that wasn't mandatory i would just skip stay home and then just play league of legends and then just before the exam i would just try to learn as much as possible and then get by with with a six which would translate to a d so yeah i would just basically aim for sixes and then be happy to just pass the exam forget everything that i that i learned and continue on to the next one and then repeat that cycle for about three years but in my final year things changed for the better because i was introduced to machine learning for the first time and text mining so these were two great courses and i wish they put them earlier in the program instead of in the final year because as i mentioned in the first year i got an introduction to python and i think there were some but i think one or two courses where you also had to use python but then you had to do group assignments and i would just get carried by uh my classmates but yeah so you didn't really have to use python throughout the degree and then in the third year we got machine learning text mining and i remember that for the first time i thought wow now i understand what i can do with this language i can make models i can automate stuff and i finally got finally got excited about learning python developing and building models and this really kick-started my data science journey and for the first time i kind of had a sense of of direction and and thought yeah i think this is where i want to go and so for my final bachelor project where i had to write a thesis i decided to do a machine learning project and this is actually quite cool so i wrote down here that this was the year that i basically fell in love with machine learning and i also published a paper so my bachelor thesis got published which is actually quite cool and if you go to scholar.google.com and search for my name then you can see this one over here detecting dutch political tweets a classifier based on voting system using supervised learning so i wrote this paper together with my supervisor so first i for the bachelor project i wrote the thesis myself and then we decided hey let's try and send it to a conference and see if we can get it published and then i did this together with with my supervisor and uh and it did it got published which is actually quite cool you can also see that it's even cited seven times by other researchers the project was basically about classifying dutch political tweets so i built a python scraper to scrape tweets that were related to politics and then we built a classifier to vote whether it would have a positive or a negative sentiment with regards to a certain party so actually quite cool yeah this is published and available online i also got to present this project at a conference in madeira which was a very cool experience and not only because i could attend the conference and present my project but also because i got to explore the island and madeira is one of the most beautiful places i've ever been so i would highly recommend checking it out if you get the chance to so yeah this is basically where things changed and as i just mentioned i wasn't particularly a star student throughout my bachelor degree but at the end i started to get some pretty good grades so for the machine learning and text mining and for my final thesis i got some pretty good grades i decided to follow up my bachelor degree with a master's degree in artificial intelligence and while picking my bachelor's degree was basically a shot in the dark as i just explained for my master's degree i had a pretty good idea of where i wanted to go with my education and later on with my career because i knew i wanted to do something with machine learning so my master degree was a two year program and this was basically the period in which i sucked a little less at python i had a better understanding of data mining and machine learning i went on an exchange to australia which was awesome and i had my first ever experience at the company working in a data science team i did the masters at the same university so at the frey and university of amsterdam also lots of courses but the main ones were data mining techniques experimental design data analysis multi-agent systems and machine learning for the quantified self those were all very specific to either working with data or data science and these courses really thought me a lot and i was also now at the point where i was interested in it and i had some some pretty okay python skills which allowed me to also complete the assignments on my own without having to rely on my classmates so this is where really where where things got going uh well for me in terms of my data science career now in the second year as i mentioned i went on an exchange to deakin university in melbourne and there i followed some parts of the data science and analytics master which was an awesome experience but most of the courses that i was taking at deakin university were covering material that i've already had in previous courses so i didn't learn that much at deakin university but that wasn't the problem at all because i was on exchange in australia i got to explore a different country and i also got to do this with two close friends of mine which was just an awesome experience and then for my final master project this was back in the netherlands i did an internship at the company and i also wrote my master thesis there my research was about a new way of building marketing attribution models using bayesian networks and i learned a ton during this internship also not only about the project itself but also this was basically the first time where i had to use uh sql to query my data instead of just using csps that i got during my courses from university but now querying from from an actual database understanding how to work with virtual machines and also basically just just team dynamics what it's what it's like to work in an actual data science team with professionals so overall a very helpful experience and that concludes my life as a student basically so um i graduated from my master's degree and this time where i mentioned that during my bachelor degree i was aiming for sixes now things changed and that was just mainly because i was interested in all the topics that we were covering in all the courses and now i managed to get really good grades i even got a lauder distinction for this degree which still is quite amazing to me if i compare that to my younger self but that is just this switch in mindset that i had where i was like okay this is what i want to do and i also started to explore and learn about things related to data science beyond just my classes so just for fun so go on youtube i would read research papers i would just look into it try and learn as much as possible and that brings us to 2019 where i started my career as a data scientist and it's actually quite a funny story so after graduating from my master's degree i tried to find a job within the data science space so i reached out to several companies also had a lot of interviews i eventually narrowed down on three interviews and also assignments for all three of them i basically got a job offer for all three of them which was amazing right out of university no experience and eventually i decided to go with the other from a large consulting company and after signing i wouldn't start until i think five or six weeks later or so and what happened within that five six week period i got an opportunity to work on a project for a company on a freelance basis so this was a was a one week project uh with a few friends of mine and i got that project through my network which was very fortunate and that small project turned out great so the company liked it i liked working on it and then i got an opportunity to continue working on it for a longer period of time and during that period i got into a serious conversation with one of the managers over there and they were like hey i want to build this team within the organization with different people with different backgrounds to basically create a small group of people that can work on new innovations for for the company and that also includes data innovations and machine learning data science innovations so i basically got an opportunity to extend my freelance project there for a longer period of time and at this point i was like okay this sounds really exciting i like what i'm currently doing but i signed a contract and i'm about to start a new job in about three weeks i think it was something like that so i basically had a decision to make so do i continue over here with a small project that could be extended or do i go with a more secure job at a big consulting firm in the end i decided to get rid of the contract so i already signed company was probably not happy with it but i signed decided to get rid of it and decided to pursue this project this freelance project to this day i'm very happy that i made that decision because it turned out great for me i'm still working as a freelance data scientist and i haven't had a single week where i didn't have work within the past three years or so so working as a freelancer that of course is is awesome because one of the major drawbacks of working as a freelancer is that you have to do outreach to get projects and it can happen of course that sometimes there are periods where you cannot work and then you also don't have an income but luckily for me that has never been the case so i'm very happy that i made that decision back in 2019. to sum this up this was actually the first time i got real world experience working on real projects so i know i mentioned i did an internship but that was still kind of my own research and this was the first time where i got to work on real projects for real business problems i also continued to learn so i make sure that almost every week i allocate time to learning because data science is such a large field and there's so much to learn and also the information changes at a very high speed so every year or so there are new models new methods that come out that beat previous models and to stay on top of everything you just have to spend some time to to stay relevant i would say i also took on some some personal projects and right now i'm currently in a phase where i'm also giving back by making videos like this and sharing the things i've learned to hopefully inspire you help you with your data science journey now as far as my career goes here on the right we have the client project to give a brief overview i've worked on data engineering projects anomaly detection models created dashboards and reports did predictive modeling also just the data analysis reports we just have to write a report and i've also helped with setting up and analyzing iot data so setting up pipelines setting up dashboards and these are highlighted in red for a reason so as you saw other red like really relevant courses i've marked them green but basically this is where it's all at i would say like eighty percent of what i've learned and what i now apply within my day-to-day job comes from these projects there's just no course that beats solving an actual business problem using data and implementing a model having said that courses are still great to start if you don't have any experience because it's pretty hard to start with a project when you don't have any experience next to my university courses i also followed some online courses so here are three so two from udemy and also this is a very good one by matt doncho from business science university this is a paid course but i think this is a really good course it's python for data science automation i also watch a lot of youtube videos which are awesome in my opinion that's also why i decided to start a youtube channel because i just love watching youtube videos so much you can get so much from them all for free and i want to do that as well also github open source projects notebooks those are excellent resources to learn and apply data science research papers a bit trickier to read to get into but the best stuff is in the research papers actually and then finally i've worked on some personal projects so created a crypto trading boat in python a personal digital twin dashboard also in python and grafana i built a personal finance dashboard and i'm currently working on this youtube channel where i also continue to learn make tutorials and develop myself as well so that basically concludes my whole data science journey from absolute beginner to now working as a freelance data scientist and as i promised i will now get into the conclusion and my 10 tips that i wish i knew when i started my data science journey let's go over here and hop into them okay so first of all there are many different ways to learn data science i've showed you my journey but this is far from ideal if we zoom out and look at all the gray boxes over here these were all courses from which i would say didn't contribute at least much to my skills my skills as a data scientist then we got a degree can help but it's definitely not necessary i think some of the online courses are two are way better than anything i got in university although i must say most jobs i think require some sort of a degree but if you have a great portfolio and you can prove your skills in that way i think you should also be fine and could easily get a job especially nowadays when data scientists are really scarce then continuing investing yourself beyond your college education this goes without saying i guess that even after you finish your initial learning and you got a job you have to continue to invest in yourself and continue to improve your skills at least in my opinion then another tip take notes and do this within some sort of digital note-taking app and try to adhere to a systemized way of taking digital notes in something like notion or obsidian or any other note-taking app that is currently available it doesn't really matter how you do it but try to build a system that is repeatable over here so when you look back that you still have those notes accessible to you so if i look back into university i didn't take much notes to be honest but some of them are in worked word documents and stuff like that i don't open those i now take all my notes in obsidian which could be a really interesting topic for another video but that is just a systemized way in which i have basically built a system for myself on how i format notes and it's all in markdown so it doesn't really matter on what device i open the notes they yeah they can continue to grow with me and that just helps so much with your learning when you can go back to to look at your old modes so that's another tip then this is a big one focus on just in time learning instead of just in case learning and what i mean by this is basically this is a more of a work from a project mindset where you take on a project and then you have to figure out how to do something and then you learn it that is just in time because you want to know how to do something and then you look it up whereas just in case learning would be oh i want to learn data science let's follow this course and what i've noticed is when you do that then during the course you're like oh this is awesome i i'm learning so much but then you're not applying the material that you've learned to tackle a project and then believe me once the course is over you will forget everything about it whereas when you do it with a project you do just in time learning and then you learn something and then do a project it sticks much better because you have translated the information that you learned within the course to your own project so that is another big one then to follow up on this one projects over courses already explained it then another big one data wrangling over building models at least in the beginning when you start out with data science i know a lot of people want to jump straight into the modeling in the algorithms that is the exciting stuff but believe me when you are working as a data scientist or data analyst 80 or even more of your time is spent working with the data to just get it into a format that can work for a model and to understand it this is where your time is spent so get really good at wrangling data and then get into building models another one learn or at least understand cloud they don't teach you this in university at least not in my courses but when you're when you're working in a job it's all cool and well that you got a nice model running within your jupyter notebook running locally but that model is of no use where you make value is when that model is implemented in production and that's all done in the cloud so try and learn something like google cloud aws or azure if you don't know which one to pick just pick one and stick with it don't try and learn uh multiple cloud platforms at once eventually when you start to work for a company they will use a certain cloud provider and you have to use that one basically but they all work kind of similar so the the the fundamentals the principles are the same then another really big one is get good at communicating and presenting results in an understandable way and i would say a mediocre data scientist that can explain uh model results in a way that the business understands it is of more value to a company than an expert data scientist that can't explain anything of what he's doing to to the business so and this just takes time takes practice but this is a large part of being a data scientist is explaining your results to the business and then to conclude you have to love what you do and this goes back to the story about my bachelor degree where i was getting sixes trying to learn python without having a real motivation to do it and understanding what i wanted to do it for and this all changed during my master's degree where i saw this trajectory in front of me where i saw myself working as a data scientist building machine learning models and that got me excited and i loved it and that is what has enabled me to stick with it over time and continue to learn to grow and to develop myself as a data scientist and that concludes my data science journey and the conclusion and tips that i gave you at the end this video turned out to be quite long if you made it all the way to the end i want to thank you i know that your attention is very valuable and i appreciate it that you've decided to give your attention to this video for this extended period i hope that you liked it that you learned something from it and that i can inspire you to continue your data science journey now if that is the case i would really appreciate it if you like this video and subscribe to the channel i'll be making more videos related to python machine learning data science basically anything to help you become better at working with data so if that's what you're interested in you should definitely subscribe i want to thank you for your time and your attention and i'll see you in the next one [Music]