The Best Computer for Data Science Beginners

Key Takeaways

hello everyone Ken here back with another video for you I get many questions about what computer to get for data science in this video I talk about the different computer hardware components as they relate to data science at the end I give you my thoughts on the best computer for the field whether that's a desktop a laptop or something else altogether as I previously mentioned one of the most frequently asked questions I get is about what is the best computer for data signs or for a beginner I wanted to make a video that gives my complete thoughts on the topic first I want to say that I'm not a real Hardware person I just know enough to be dangerous I've done a few PC builds and I may or may not have a video on one coming up in the next few months please take my hardware knowledge with a grain of salt that's all I kind of wanted to get out of the way first let's start with the CPU here this is important for doing most of the basic processing functions for data science in 2020 better processors usually have more cores this allows more efficient parallel processing most data science packages are run using hardware from the CPU resources generally you can get by with a very basic processor but having a powerful multi-threaded CPU can potentially speed up your workflow for a desktop build I personally use the AMD rising processors I think the Rison 5 7 and 9 are the most practical for a data science build there's nothing wrong with the Intel processors I just personally prefer Andy I think they're cheaper and you get just as good a product next comes the RAM which stands for random access memory in simplest terms the amount of RAM that you have dictates how much data you can process at one time in this sense more ram equals larger batches and quicker processing following that still memory is the hard drive this is likely where your data will be stored I generally don't think that you have to go crazy here there are plenty of cloud storage providers that are very cheap you can just download your data locally for a short period of time if needed I think 256 gigabytes of hard drive memory is plenty this type of memory is cheap so if you want to scale up feel free to it won't cost you that much I recommend going with an SSD which stands for a steady-state Drive this is far quicker for data retrieval it isn't as fast as RAM by a longshot but it can be an order of magnitude faster than the traditional hhd disk drive finally the graphics card can also be extremely important for deep learning tasks I think a GPU with around 6 gigabytes of memory will meet most needs but if you're working on huge datasets editing videos or playing graphics intensive games you might want to invest in a more expensive graphics card here I recently received this one from Nvidia and I'm extremely excited to experiment with it more potential videos to come on that in the future now for my final computer recommendation with all this being said the computer do you have or one that you can afford is the best computer for data science with the advancements in cloud computing most data science isn't actually done on your local machine anymore you have access to top-of-the-line GPUs like this one for free or for extremely cheap on these online platforms Google collab and Kaggle allow you to access CPU and GPU even TPU compute power through your browsers one thing I didn't actually talk about was internet connectivity that may actually be the most important feature of a computer for starting data science if you still want to do things locally you have additional GPU options there's an upcoming trend of EGP use that allow you to connect powerful graphics cards to your laptop via USB C this allows you to potentially connect very powerful and video graphics cards to your laptop and leverage their power there are also tools like Intel's neural compute stick that offer similar portable compute power for neural nets and computer vision these are still relatively new and you have to make sure that you have the proper requirements and configurations on your computer to make them work so definitely do your homework before exploring these options so for beginners I wouldn't worry that much about the hardware you can do literally anything you need to do on a computer that is under $500 still there are use cases for a more advanced can and if you're trying to build one or certain that you want to be working locally I recommend the following specs for a CPU you generally want to have at least four cores again I think that AMD Rison five through nine lines are perfectly fine for a data science PC for RIM you generally want to have at the very very minimum eight gigabytes you probably want to start around sixteen gigabytes and go any you know go as high as you want there for hard drive I've mentioned that 256 gigabytes is enough if you want again this memory is not that expensive so you can ramp up to even like a terabyte I believe this was around $150 which really isn't that bad for a GPU it really depends on your use case if you're doing a lot of video editing or or playing high-performance games you might want to invest a little bit more here if you're only using it for deep learning that frankly the online computer resources can be just as cost effective I think that almost any computer that has these specs will suffice I personally been using a Dell XPS 15 inch model for the past few years as my laptop and I've built my own computer for a desktop the desktop has quite a lot of upgrades because I use it for video editing as well as data science I've linked all the parts for my personal PC my laptop and the computer that I'm building in the description below definitely those are really top of the line requirements they're not something you need to get started so I've also included a kind of mid tier level data science computer and a cheaper tier data science computer that can be built for you know less than a thousand dollars I'm actually in the process of building an absolute beast of the computer for data science so stay tuned for the video on that in the future I hope that this video helps you to understand more about the hardware requirements of data science and also about the power of cloud computing thank you so much for watching and good luck on your data science journey