M1 vs Intel Mac for Data Science

what up dad nerds i'm luke and i'm a data analyst i recently got the new macbook pro with an m1 chip and i'm going through and setting it up for all the tools i need but i thought i'd share a lot of the lessons learned to you as well specifically i wanted to look at how it performs for three basic types of operations for data science one being spreadsheets so things like microsoft excel two being coding so common ones that i use of python and r and then finally business intelligence tools such as tableau and power bi so to start let's talk about some basic specs for my new m1 macbook so the new macbook is a 13-inch macbook pro and it has 16 gigabytes of memory and it's using the m1 chip and from the research i've found is that the m1 chip performs very similar between the macbook pro and also the macbook air so a lot of the stuff that i'm doing in this video can also be applied to the macbook air additionally i'll be performing some performance tests comparing my new m1 mac to my previous intel mac my previous intel mac was a 13-inch macbook pro with 16 gigabytes of memory and an intel i5 processor that i got back in 2018. okay so let's jump into what tools are available for spreadsheet users first tool we're going to be looking at is apple's numbers so this is an app that is native to the mac and so because of that it's going to be compatible it has everything that you may need for basic spreadsheet operations and just a little fun fact you can also import and export excel files for it so if you need to in a crunch and you don't have microsoft excel you could use numbers for this the next tool to look at for spreadsheets is google sheets and there's actually no app that you could actually get to install for this application like you can on your phone or an ipad instead it's native to your web browser so i use google chrome but that's how you would access it and because you're using it via your browser it's gonna have the same compatibility regardless of the architecture okay so let's move in the tool that most people actually use for spreadsheet and that's microsoft excel and this one we're actually going to be doing more of a deep dive for analysis so i went in and i downloaded microsoft office suite i have a license for it and i was able to download it and went and installed very simply and very quickly actually what i thought was really interesting with this was microsoft has actually upgraded all their products to run natively on apple silicon so you don't have to use rosetta for this so now that i have microsoft excel installed i wanted to go through and see how it performed compared to my old intel mac specifically i just wanted to launch the application do some common operations such as copying pasting entering formulas dragging those formulas down and then i wanted to perform a test as well and the test that i performed was i had a list of a million rows and they had random numbers within it and i wanted to use the filter function to sort that row or sort yeah column of a million numbers and see how long it took so what i found was that the new m1 mac actually sorted the row of about a million values in a much faster time so it took about five seconds to sort this row sort this column and where my intel mac took about nine seconds to additionally during this process of just messing around with it excel further i found that excel was so much more responsive on the m1 mac and so much quicker compared to my previous intel mac where i'd get the spinny little circle saying that it was trying to load so overall i was very impressed with performance of microsoft excel on the m1 macbook so for those spreadsheet users out there if you're looking to get the new m1 macbook for spreadsheet operations i highly recommend it you have a multitude of different tools to actually use and microsoft excel operates even quicker than the old intel max okay so let's move now into coding on an m1 macbook we'll be looking at two specific languages we'll be looking at r and also python for data science there are two primary ways of getting python on your computer you can either use python.org or the anaconda distribution to load a package of different tools that also includes python for python.org whenever the new version of python so python 3.9 it is supported on apple silicon and that's really awesome if you want to use it for that but a lot of the libraries and packages are still under development and working out a little bit of kinks so i wouldn't necessarily recommend using uh this approach just yet on the m1 macbook for your data science needs i'm going to be using the anaconda distribution and it hasn't been updated yet for the apple silicon architecture so it's going to be run through rosetta but i feel like this is going to accomplish more than enough of my need so let's start looking at it through using python through the anaconda distribution so for this i went to the anaconda website and downloaded the installer for anaconda and as you can see in this example whenever i download spotify which also was a application that didn't run natively on apple silicon it prompts you to accept and use rosetta and so i did that also during this time when loading the anaconda distribution since we're running it in the intel architecture using rosetta all packages are supported so i went through and just just a test but tested you know numpy sci-fi pandas and scikit-learn and tensorflow and i didn't have any issues when actually running through and operating it uh for these different packages they all worked perfectly fine finally for python i wanted to go in and test how long it takes the python code to operate on my new macbook compared to my old intel mac and for this what i wanted to do was just do a simple operation that i could use a timing to time how long it takes to append a list of numbers uh from one to ten thousand so when i ran this with my m1 macbook this took about 17 seconds and when running it with my intel macbook my old one took about 22 seconds so overall what i found was the performance on the new m1 macbook was a 20 increase in the performance for running uh python code and additionally whenever i was just going through and using python in general on this computer i felt it wasn't uh very clunky at all it was operating just fine but overall i didn't find a lot of big differences between my old mac and my new back and then my new mac so if you're hardcore python coder and looking to get the m1 macbook for python i highly recommend it it seems fully supported especially if using the anaconda distribution if you're really keen on sticking with that python.org so that core python only i would maybe wait a little bit till it gets further developed so next let's look at another popular coding language and that is r and there's two versions of r that you could get to whether you want to install it using rosetta and intel architecture or using it on apple silicon so the newer version of r is available uh for a beta to run on apple silicon but like i said it's in the beta similar to our python i wouldn't necessarily recommend this at the time they do have a stable version that runs via rosetta and that's the one that i used and we're going to go through and actually test that version so for this i went in and installed r and then i don't really know how to use r through the command line so i installed rstudio it installs uh hosted in rosetta using intel as the architecture similar to python since it's running in this intel architecture all the packages are supported so specifically i'll just call it ggplot tidyverse and shiny all of them i had no problems loading and actually operating within rstudio from there i wanted to do a performance test once again comparing my new m1 macbook to my old intel mac and i want to just do a similar test that i did previously with python so joining numbers 1 through 10 000 in a string and seeing how long that it takes when i ran this on my new m1 macbook it came out in 1.4 seconds-ish and for the old intel mac it was about 2.4 seconds-ish and so that's about a i mean 40 improvement uh for the seconds wise not super great improvement but it is an improvement nonetheless so overall what i found with r similar to python uh very responsive operated super well within the m1 macbook and so if you're an r user i highly recommend it get the m1 macbook and use it and eventually they're going to update r as well to be supported on apple silicon okay the last segment we're going to be looking at business intelligence or bi applications that you can use and specifically we'll be looking at google data studio tableau and also power bi so the first one we're going to be looking at google data studio for this google data studio doesn't have an application per se it's going to be used within your web browser so like i was mentioning previously i use google chrome and i found when operating within google chrome operating just fine as expected so if you're a google data studio uh studio guru and you're using that a lot i don't see any problem with continuing to use it on an m1 macbook okay so moving into tableau this is when the bad news starts to happen with m1 macbook so whenever i went through and tried to go in and download and install the current version of tableau onto my computer i ran into some issues and so i had to research this further and upon research i've found that the current version of tableau so 2020.4 it's not currently compatible with apple silicon i did find that you can do a work around and if you needed to get tableau on your computer you can install a previous version so i did that i installed uh version 2020.3 but from what i've read so far i don't think it's going to be necessarily fully compatible or even fully supported by tableau so because of these different issues yes if you already have an m1 macbook and you need to use it for tableau you can do it but i don't necessarily i wouldn't recommend it right now to upgrade your new mac to run tableau on it because tableau isn't formally supporting it right now okay and the final bi tool to cover is power bi and you're probably saying luke power bi is a microsoft based application how are you even installing it on an apple computer and to that i say i run a virtual machine on my old intel mac and through that i can host windows within it and i'm able to actually use power bi in the virtual machine if i need to it's a little slow and a little clunky but overall it can get the job done one other option to note for windows on a mac computer previously on the intel max you could use bootcamp to do a dual boot whether you wanted to launch windows or you wanted to launch apple's operating system but unfortunately bootcamp on the new m1 macbook isn't supported so you can't run windows via bootcamp on the macbook additionally i looked into different virtualization machines so different vms to be hosted on the m1 macbook and all of the vms are currently under development and they're still trying to update the the coding so that way those virtual machines can actually run on this new apple silicone chip they do have a one option that works right now but it's quite lengthy on how to install and i'll include a link in the description so if you're looking to go that route to install a virtual machine check it out but it was just overall too complicated and too burdensome so i just abandoned it so for power bi i would wait uh if you use it on your mac i would wait until apple and microsoft develop further to support microsoft on an m1 mac so just a quick recap if you're a spreadsheet user you have a multitude of different options you can use microsoft excel google sheets or apple's numbers and all of them are fully supported on the apple silicone so highly recommend switching to the m1 macbook if you use it primarily for those operations for coding we tested out python and r both of them have beta versions that you can run on apple silicon but they i recommend running them via intel using rosetta so my recommendation for those coders out there that are using python are definitely check out the m1 macbook for this and then finally business intelligence tools so google data studio works in chrome browser it's a no-brainer you can continue to use it things like tableau is not yet supported and then things like virtual machines and bootcamp don't support microsoft being run on your macbook so for business intelligence users i don't really recommend getting the m1 macbook right now it's not mature enough with the software development from third parties and so you're gonna need to wait a little bit for this so with that i hope you got some value out of this video smash that like button to help me out and with that let's run some b-roll footage of my new m1 mac [Music] you

Channel: Luke Barousse

Views: 54,623

Rating: 4.8923283 out of 5

Keywords: Macbook for Python, Macbook for R, Macbook for Tableau, Macbook for Power BI, Macbook for Business Intelligence, M1 for Business Intelligence, M1 for BI, M1 for Python, M1 for R, M1 for Tableau, M1 for VMs, M1 for Virtual Machines

Id: vyPm2fOyS7Y

Length: 15min 4sec (904 seconds)

Published: Fri Jan 01 2021