Mac vs PC for Data Science (SPOILER: Don't get a Mac...)

hey what's up dad nerd luke you know i don't like one yeah call you that hey can you airdrop me that python file that you did for boss where you did that papaya marketing analysis i need to steal some code from it um airdrop uh i can email it to you uh nevermind um i forgot you weren't in the apple ecosystem email isn't that hard uh yeah it is i have to wait for your email download the attachment move the attachment to where i want it from there clear my inbox just to get a code to snippet that i need just forget it i'm gonna go to stack overflow instead um okay and just got it thanks nerd hey luke yeah what's up do you want to go to happy hour with me and the new data engineer i'm about to start a group chat in order to share all the details uh thanks but no thanks i can't be in a group test message with somebody that causes those green messages green text messages yeah you have that barbaric android phone those green messages ruin group chats especially since i use my computer to text from you get text messages on your computer yeah welcome to 2021 my dude um bye later [Music] what up dead nerds um real luke here and i'm a data analyst and my channel is all about tech and skills for data science and one of the most common questions i get about computers is whether or not to get a mac for data science so as shown in that opening skit there's a lot of peer pressure from others and myself to pressure others into entering this apple ecosystem i myself love apple products and i use them on a daily basis but when it comes to macs and my macbook this wasn't necessarily designed for data science and sometimes we have to go about installing workarounds in order to get the applications we may need take for example my favorite application so power bi this doesn't run natively in mac os so i have to install a workaround of a virtual machine in order to get this application to work on my mac so obviously i'm pretty biased towards mac computers and in order not to have that confirmation bias if you will i set out to youtube and posed you with a poll of what is your operating system of choice for data science with over four thousand respondents it's pretty clear that the most popular operating system chosen for data science is windows followed by mac then linux and finally a combination of operating systems looking at the numbers more simply we find that for every two data nerds that use windows there is one data nerd using mac and half a data nerd each using linux and multiple operating systems which when looking at it from this perspective there actually doesn't seem like much of a difference with that aside i do think that this poll does a great point in showing what is the dominant operating system for data science even though it pains me to say as a mac fanboy that windows is this dominant operating system i feel that beginners in data science really need to strongly consider this windows machine as their first choice so this begs the question why are windows machines the better options i find for two main reasons first is the native application support of most all data science tools although there are a lot of popular options for both windows and mac including programming languages such as python and r or spreadsheets and sql tools when you start getting into more niche tools the availability of tools across cross operating systems becomes limited but you can typically find that windows is almost always supported for example in my day job as a data analyst i've run into problems with these applications and that they've only been supported on a windows operating system now i will add one disclaimer in that a lot of these applications that are specific to windows only are based in gears towards large corporations in that case you're going to be typically working for a company in which they'll give you a work computer and so having windows in that case is not going to be a problem um editor luke here actually second disclaimer i realized that previous advice that i gave was more geared towards data analyst when it comes to people like data scientists and also data engineers a lot of the work that they do is going to involve a remote machine or even in the cloud so the work they're going to do it's going to be accessing it through the web browser even the command line so in that case an operating system it's not really going to matter which one you select so with that back to regular luke the second and final reason on why windows machines are the more popular option has to deal with the lower cost of these machines browsing amazon you can typically find a hefty enough computer in order to handle your data science workload needs for around 500 to 1000 bucks so let's better show this by looking at some of my recommended options first is the asus vivo book which at only 450 dollars is a pretty great deal especially since it has 8 gigabytes of ram next is the lenovo ideapad for around 700 bucks i feel like this is actually the best value as it has a pretty high amount of ram with eight gigabytes and also a pretty good gpu with nvidia so that way you can do some sort of machine learning with it finally we'll wrap these low cost options up with the acer nitro 5 which is around 800 bucks and also has about 8 gigabytes of ram and a nvidia gpu now if you're considering another computer besides this i think that's perfectly fine i think you should just mainly consider three main specs first is memory i recommend a minimum of eight gigabytes if you're going to be doing things like virtual machines then you need to be considering 16 gigabytes or higher for the cpu if it has an intel processor i recommend an i5 or higher although the i3 may work as well and then for the gpu if you're going to get into machine learning i highly recommend that you get something with a nvidia graphics card in order to take advantage of cutter processing if you're looking for a computer more in the thousand dollar range i highly recommend that you check out the dell xps series i have an xps 13 that i use as a backup to my mac and i've been using it for data science and haven't run into really any issues and feel it's a hell of a computer so wrapping this section up on why you should consider a windows machine i consider it a safer choice especially those for those newer to data science based on its cost and its compatibility with apps and data science now let's say that didn't convince you and you're still going to go with or continue to use a mac so going back to that youtube poll that i created i was curious to not only compare what operating system people are using for data science but also in the general market so what normal people are using for their operating systems so i built this clustered column chart that compares overall market share to data science market share for the three top operating systems for desktop and laptop computers 88 of users choose windows as their operating system of choice but when you look at it for data science from my poll it's only 56 percent now looking at mac os the number of users that use this is almost three times higher than what the general market uses of 10 and when we look at linux it's even more so taking this analysis a step further when comparing the likelihood of selecting an operating system for data science over that of the market we can see that the mac and more notably linux are more likely to be selected so personally what i feel is happening is that there's being a shift to mac and also linux machines in order to be used for data science so if you decide to go this option i think you're in good company diving into the application support on mac os i feel that mac os may potentially be good enough for most beginners i made a video last week detailing what applications do work for mac os and what applications require a workaround to work i do want to call out real quick the workaround that i use to get windows on my mac os and that's parallels and that's what i've been using actually for all of my analysis in power bi and i've really ran into no issues using power bi on my mac if you're interested in trying out parallels i have a free 14-day trial in the description below but even with some applications requiring that windows operating system to work i'm seeing that many data science applications are now shifting to the browser to be operating system agnostic so if you're still convinced and you're in the market for a mac machine my number one choice right now is the macbook air it's on amazon right now for 899 bucks and i think that's a hell of a deal i've tested the m1 chip and feel like it's more than sufficient for data science tasks so i don't recommend them m1 max or pro for everyday data nerds if you're going to upgrade anything in the computer i highly recommend that you consider upgrading the memory to 16 gigabytes as this will be better in supporting multiple applications virtual machines and even containers but if you're only doing routine data science tasks such as excel tableau or python eight gigabytes is more than enough as always if you got value out of this video smash that like button with that i'll see you in the next one [Music] you

Channel: Luke Barousse

Views: 201,187

Keywords: data viz by luke, business intelligence, data science, bi, computer science, data nerd, data analyst, data scientist, how to, data project, data analytics

Id: U4vh2EClJic

Length: 9min 36sec (576 seconds)

Published: Thu Dec 16 2021