5 THINGS I WISH I KNEW Before Starting Data Science

data science has just been named the industry of the century but what does it take to get started let's find out more hmm data science tell me more what's happening guys my name is nicholas renate and in this video we're going to be taking a look at five things i wish i knew before i started data science so these are things i wish i knew before i started on my journey to becoming a full-fledged data scientist things that would have saved me a heap of time and made me a whole heap more efficient along the way ready to do it let's get to it tip number one pick a language and stick with it when i first started out on my journey as a data scientist i tried a whole bunch of different languages i tried ah i tried octave i tried javascript scala julia a whole bunch of different languages now this basically meant that i spent a ton of time just spinning my wheels so rather than picking one and getting really good at it i basically spun my wheels and stayed in beginner land rather than progressing further forward and progressing towards becoming a senior data scientist so picking one language is going to help you avoid making that mistake now as a general rule of thumb you want to pick one programming language and ideally something that's going to serve you in the longer term now i may cop a little bit of hate for this but my personal advice is to pick python and get really really good at it now the reason why i say that is because there's a whole heap of stuff you can do with python plus a lot of the advancements in deep learning are happening in python first so you want to be able to leverage that now but also in the future in terms of how to get started learning python i'd highly recommend you check out the python crash course book by eric mathis so it's published by no such press and it's what i use to begin learning python and it'll help you get a kickstart a whole heap faster tip number two as the title suggests being a data scientist means you're going to work with data quite a fair bit now what a lot of people don't recognize is that you're going to be spending up to 80 percent of your time actually just pre-processing and making sure you've got good data now good data might mean you've got appropriate data quality you've collected enough data you've collected data that doesn't have bias or you understand the bias but it also means pre-processing the data and working with it to get it into a state that you can start modeling in becoming really good at processing data is going to make your life a whole heap easier when it actually gets to the point where you need to develop algorithms with it now in terms of how to work with data if you're starting out from tip one and learning python i'd highly recommend you become really familiar with the packages numpy pandas and opencv if you're going to be working with images now the best way to take a look at how to learn some of these packages there's a whole heap of free videos and free courses out there i've also created a couple if you want to check those out no pressure tip number three learn to present your ideas outside of a jupiter notebook now as data scientists we tend to want to work and write code but you've got to keep in mind that the audience that you're presenting your ideas to aren't familiar with code and probably don't want to get involved in code so presenting your ideas in a format that's familiar to them is going to make them more receptive to taking on what you have to say also presenting your ideas in plain english rather than speaking in really complex or technical terms is going to make them more interested and really establish you as a trusted advisor so the next time there's something that's data sciencey they'll go and know that hey that guy or that girl knows how to present this in a way that i understand so i'm gonna go to them to get some advice now presenting your ideas in a format outside of a jupiter is really simple right all you have to do is create a powerpoint copy some of your diagrams into that and that's a great start tip number four learn about the industry that you're working in this tip cannot be emphasized enough so learning about your industry is so critical whenever you're working in data science because you want to know what factors are likely to influence a particular outcome so say for example you're building a forecast model for a retail company well you want to know a little bit about the industry that that particular company is working in because that's going to influence seasonality trends and what particular events that happen in a calendar year that are likely to influence your model by knowing these things you're able to cater for them when you're actually pre-processing and working with your data likewise say you're building a computer vision model that's detecting lesions on your skin well if you're well-versed as to what a lesion actually is or what it actually looks like when you're labeling your data you're going to be able to more easily identify what is a true lesion versus what isn't so having industry knowledge goes a really long way now if you don't happen to work in the industry that you're currently operating in a great way to get a head start is to read company financial reports but also take a look at ibis world so they've got some really great industry reports on what's happening with that industry what processes look like now it is a paid subscription but more often than not you will be able to access this through your university tip number five now in order to really stand out as a data scientist kaggle is a great place to start that being said there's a lot of data scientists competing for attention on kaggle so in order to stand out you got to do something a little bit different now the best way to do this is to go and find a passion project something that you're really interested in now take that and start building a data science use case around it so in my particular case i was really interested in what we could potentially do with sign language and data science or ai so i started building a sign language recognition model that would allow me to try to detect different poses now admittedly it's still really really basic but it's something that i'm passionate about and it allows me to show what types of capabilities that i've got in my toolbox finding a use case like this is so critically important because it really allows you to stand out from the pack now you're probably thinking well nick how do i actually go and showcase what have gone and done well a great way to do this is actually by setting up a github profile and making sure you have a great readme so i'll include a link to the description as to what a great readme actually looks like but basically you want to be able to walk each user through how do you use your library or how to use your project this is going to help you stand out head and shoulders above when you're actually becoming a data scientist tip number six now i know that i said that there are only going to be five but i thought this bonus tip was so so useful one of my favorite books is atomic habits by james clear and the whole premise of the book is setting small habits that you can do every day i think setting an atomic habit and dedicating five minutes each morning to writing a couple of lines of code or investigating something new in data science is going to make you a whole heap better a whole heap faster so by doing this each and every morning you're going to compound your skills and get a whole heap better a whole heap faster so as soon as i get up each morning i make a coffee and i spend five minutes actually investigating something new in data science or writing a particular line of code or a couple of lines of code that i'm really unfamiliar with this helps me push myself outside of my boundaries and learn a little bit a little bit faster and that about wraps it up thanks so much for tuning in guys hopefully you found this video useful if you did be sure to give it a thumbs up hit subscribe and tick that bell so you get notified of when i release future videos and let me know of any tips you had when you were getting started in data science thanks again for tuning in peace

Channel: Nicholas Renotte

Views: 29,559

Rating: 4.974299 out of 5

Keywords: things I wish I knew before starting data science, things i wish i knew, data science, data scientist, data science career, 5 things i wish i knew before becoming a data scientist

Id: B5emClQf_I4

Length: 7min 32sec (452 seconds)

Published: Thu Nov 26 2020