The 7 Biggest Data Science Beginner Mistakes

Key Takeaways

as crazy as this might seem at one point in time i too was a data science beginner i've come a long way but it hasn't been without my share of mistakes in this video i'll walk through some of the beginner mistakes that i made when i was learning data science and some of the ones that i've seen other people make as well hopefully this video helps you to avoid the common pitfalls of learning this field make sure you stayed at the end because we finished on a positive note with some ways to avoid these traps the first mistake that i see people make is that they spend too much time deliberating on where to start people are constantly asking me which resources are the best and which they should personally pursue i get it i 100 faced this exact same problem when i was beginning my own journey but over time what i found was that almost every resource that i've tried and you know the ones that are fairly widely used they're pretty darn good the teaching styles of the different courses that are out there are a little bit different but almost every resource also has some free material so you can experiment and see if it matches your learning style my advice is to explore a bit find one that you kind of like and dive right in don't spend time worrying about if you made the wrong choice or not something that people don't realize is that just because you learned from one course doesn't mean that you shouldn't also take other courses in the future and learn more information learn the same thing from a different angle learning is additive and learning from multiple places is encouraged and can help expand your knowledge there's no one right place to start for anyone and you should find a resource that seems good enough and just jump in either way the only way you'll truly know it's wrong is if you try it and you don't like it [Music] i'm a firm believer that most people learn best from examples unfortunately when learning data science it's really easy to get wrapped up in the theory and not be able to connect it to the real world i've seen so many new data science learners say that they know all the theory behind linear regression but they're completely dumbfounded when it comes time to apply that knowledge don't get me wrong the theory is very important but you can actually get a better understanding of the theory if you're applying these models in the real world i firmly recommend that you start getting involved with data science projects far before you feel like you're ready getting involved can be as easy as just reviewing other people's projects on kaggle or watching the project examples that i've put together on my youtube channel something that scares a lot of people is that they think a data science project has to be something big and scary a project doesn't need to be some end-to-end data science web application or you've also gone and scraped the data it can be something as simple as writing a for loop to print your name over and over again in my mind a project is anything that's hands-on where you're applying your knowledge and you're also creating an output start with very small projects little toys and advance to bigger and better things also these big projects can be broken down into very little tiny projects and each of those individual small projects aren't very overwhelming to start some of the projects that i've done on this channel have been scraping data science job applications and analyzing them i've broken that into seven parts making a sports ball classifier and creating a youtube leaderboard for everyone who comments on my videos through these i've broadened my knowledge of git of web scraping of computer vision and using tools like docker i don't think if i studied any of these concepts or tools independently i'd have as good a grasp of them as i do now one small project that would really help me out is if you like this video subscribed and turned on notifications it helps us to all beat the youtube algorithm together [Music] every week i probably get around 20 people that ask me for individual mentorship while i'm extremely flattered at least right now i don't offer this type of help to people while i think overall mentorship is a really good thing i also believe that it should be reserved for the later stages of the data science journey one of the fairly unique things about data science is that often there isn't a single person that has all of the answers i find that it helps to crowdsource or go to a community to solve the problems that you're facing as an individual especially in the early stages i also think that it's really important to become self-sufficient when you're starting out i would say that almost every beginner technical question has already been answered online somewhere learning where to look and to debug your own work is one of the most important skills that you can learn in this field trust me you'll be doing an absurd amount of debugging for any project you work on it might be a bit of tough love but struggling with the early problems is a very important learning experience for any prospective data scientist continuing on from the previous mistake i find that most beginner data scientists give up on learning just a little bit too early when things start getting hard or when people get stuck it's really easy to stop the process i was 100 someone who fell into this category but after reading about one zillion self-help books i found that if i push through this resistance that is where the true learning really took place it's proven that we learn the most when we stretch our capabilities and we almost always have more in the tank than we think actually one of my favorite i guess author's personalities whoever it is david goggins thinks we have about a hundred percent more to give when we think we're completely done yeah your hands stay hard i personally have a ten minute rule that helps me with this when i feel like i'm done i force myself to go on for another 10 minutes if after that i want to stop i can but i usually find myself catching a second wind and gaining momentum and really not wanting to stop anymore because i feel like i'm learning so much [Music] so this might sound contrary to the previous point but please bear with me different people will face different obstacles when they're learning and i'm trying to cover my bases as much as possible some people actually try to learn too much too fast and they completely burn out data science is a marathon and not a sprint in this career you'll always have to keep learning and it just isn't sustainable to try to learn everything in a very short period of time from observation the people who make it in this profession are the ones that love what they're doing and are slowly chipping away at the gigantic body of knowledge to be fair there are some people that catch an obsession with the field and really don't burn out after long hours if you're one of these people great but i haven't seen too many people with these characteristics that can keep that fire of obsession going for an extended period of time for this i recommend again creating good habits creating systems where you're learning data science continually don't worry about all the knowledge that you have to learn just focus on what you can learn in an individual day or in an individual study session so i touched on this a little in my video on why you probably won't become a data scientist but a big thing that many beginners overlook is that data science might just not be a good fit for them data science is an appealing career because of the income and the cool problems that you get to face but the reality of the work often isn't that exciting for most data scientists you spend a lot of time coding a lot of time cleaning data and not a lot of time working on sexy deep learning and computer vision problems i think you have to really enjoy the work or have an extremely compelling purpose that motivates you to do these types of tasks and have longevity in this career i would take some deep introspection and think about if this is a really good fit for you or if you're just kind of getting caught up in the flashiness and the excitement of the career when you ask someone what makes a good data scientist most people tell you that they have to be good at statistics or programming what most of the population won't tell you is that you also have to have a very strong understanding of the topic of your analysis knowledge the true value of data science work comes from creating solutions to real world problems i would say a good data scientist is someone who can drive value with their work regardless of how fancy their code is and how complex the algorithms that they use are this is one of the reasons that i recommend that you work on projects related to things you're passionate about you can show your subject area expertise through these and really highlight your ability to go deep on a topic my specialized experience in sports is something that has really carried me a long way and created many opportunities for me on my podcast ken's nearest neighbors i also really try to highlight the ways that my guests have differentiated themselves from the pack notice that these mistakes don't focus on the order of learning things the programming languages that are needed or the tools i did this on purpose in data science tools change the content you learn can also change depending on your specific job i believe it's more important to focus on avoiding the behavioral obstacles over the technical ones these are a few ways that you can overcome these in your life the first thing i recommend is to create great habits much of the data science mistakes can be addressed by creating systems that create positive feedback for yourself i created the 66 days of data and the community associated with it to help you create good habits and have a place to learn with others the second thing i recommend is to have a bias towards action too much deliberation really leads to stress taking action in your learning and in your projects helps you to build momentum as well when you learn something turn around and apply it immediately and remember that projects don't have to be these big scary things they can be small toy examples that you're just having fun and experimenting with the third thing i recommend is be real with yourself and your expectations some people can learn enough data science in six months to land a job i was not one of those people some people can sit and code for five hours straight again i'm not someone who enjoys long hauls like that experiment with your own personal boundaries and match your learning efforts to them data science isn't a competition it's a long journey of slow and enjoyable progression work at your own pace and i assure you that you'll enjoy the ride a lot more also if you find the data science isn't for you that's totally okay there are plenty of other professions that aren't as heavily focused on continued learning with so much overhead associated with upskilling i hope you enjoyed this video thank you so much for watching and good luck on your data science journey [Music]