StatQuest: K-means clustering

K-means clustering is used in all kinds of situations and it's crazy simple. The R code is on the StatQuest GitHub: github.com/StatQuest/k_means_...
For a complete index of all the StatQuest videos, check out:
statquest.org/video-index/
If you'd like to support StatQuest, please consider...
Buying The StatQuest Illustrated Guide to Machine Learning!!!
PDF - statquest.gumroad.com/l/wvtmc
Paperback - www.amazon.com/dp/B09ZCKR4H6
Kindle eBook - www.amazon.com/dp/B09ZG79HXC
Patreon: / statquest
...or...
KZbin Membership: / @statquest
...a cool StatQuest t-shirt or sweatshirt:
shop.spreadshirt.com/statques...
...buying one or two of my songs (or go large and get a whole album!)
joshuastarmer.bandcamp.com/
...or just donating to StatQuest!
www.paypal.me/statquest
Lastly, if you want to keep up with me as I research and create new StatQuests, follow me on twitter:
/ joshuastarmer
0:00 Awesome song and introduction
0:33 The K-means clustering algorithm
4:26 How to pick a value for K (How to use an elbow plot)
6:06 K-means vs Hierarchical Clustering
6:28 K-means clustering and 2-Dimensional data
7:08 K-means clustering and heatmaps
Corrections:
5:58 I should have put "Reduction in Variation" instead of "Reduction is Variation"
7:25 Point (7,-8) should be in the lower right-hand quadrant.
8:25 The "nclust =25" in the arguments is actually "nstart=25"
#statquest #ML

[@statquest]

Corrections: 5:58 I should have put "Reduction in Variation" instead of "Reduction is Variation" 7:25 Point (7,-8) should be in the lower right-hand quadrant. 8:25 The "nclust =25" in the arguments is actually "nstart=25" Support StatQuest by buying my book The StatQuest Illustrated Guide to Machine Learning or a Study Guide or Merch!!! statquest.org/statquest-store/

[@sherlingchan5054]

Can you explain how Reduction in Variation is calculated? [4:09] I am guessing it is 1 - (the length of the 2nd cluster attempt/ the longest length of all attempts)? Is it correct?

[@statquest]

@@sherlingchan5054 It depends on the distance metric. In this case, I was just adding up the variances for each cluster and picking the one with the smallest value.

[@mriganktiwari3782]

Did you also mean dimensions instead of samples, while explaining distance calculation? just before the hooray part ?

[@statquest]

@@mriganktiwari3782 The "axes" that I mention are synonymous for "dimensions". In this case, the number of samples "samples" is the same as the number of dimensions.

[@Jkfgjfgjfkjg]

@6:40 You have it spelled both "Euclidean" and "Euclidian".

[@gatsbyliu1084]

Who is watching this 1 day before the final exams?

[@statquest]

bam! :)

[@jagjyotsingh8456]

😂😂 love from University of Maryland’s MBA cohort!

[@ApheliosLover-hr8cd]

Midterm exams *

[@mahfuzrahman7215]

On my way to final😂😂

[@odayhigaze8302]

Bro im watching this 2 hours before the exam , i havent sleep all night

[@DescomplicandoMusica]

You know what? I'm studying machine learning for a while...I watched videos from various channels and bought courses. But you're incomparably better than all of them put together. Being didactic is a talent, now that I have found you I will watch all your videos. By the way, if you have any deep learning course, I buy them right now, no matter the price. Thank you very much for your work!

[@statquest]

Thank you so much! Muito Obrigado! I really appreciate it. :)

[@statquest]

I haven't don that before, but I'll think about it.

[@statquest]

@@DescomplicandoMusica Will do.

[@developerfoe]

@@statquest i'll buy too...please make a course on deep learning.. i am a huge fan of you..

[@statquest]

@@developerfoe thank you!

[@piyushraj5171]

The level of clarity I have achieved watching your videos is unparalleled comparing from other resources. Thank you so much for your awesome videos.

[@willd0g]

Definitely agree

[@wolfisraging]

Best teacher in the world

[@statquest]

Thanks! :)

[@vladimirbosinceanu5778]

I would have to agree.

[@sparkartsdistinctions1257]

Thankyou 😌

[@fxsurgeon1]

BAM!

[@shubhlaabh6342]

@Wolf Rage Absolutely

[@PorthoGamesBR]

Your videos are so good! They are calm, without being boring, in the exact rythm to understand the concept without getting tired of it, and doing it so perfectly is an art. Congratulations StatQuest, you got another sub!

[@statquest]

Thank you!

[@michaelyang7657]

He explained everything so simply and clearly, a very solid tutorial video to start with!

[@statquest]

Thank you very much! :)

[@giannigoethals5656]

I'm having Business Intelligence this year and you are just clarifying every single piece that wasn't clear, thank you so much!

[@statquest]

Glad it was helpful!

[@simonroques8579]

Just...Woaw ! I was looking for simply explained KMC and I found StatQuest, the best videos ever (on statistics) !

[@maxprotube5665]

While trying to comprehend K-means Clustering by watching other videos, I struggled to grasp the concept of selecting centroids and assigning them to specific groups. However, this particular video was extremely helpful in clarifying everything for me. I am grateful for the assistance!

[@statquest]

Glad it helped!

[@Derrekito]

You explained in just over 8 minutes what a 70-page research could not. Thank you so much.

[@statquest]

Thanks!

[@diversehorizons]

I really like the way you visualize and slowly explain the most difficult concepts. It is very helpful. Please make as many videos as you can! I think you're on to a good business niche.

[@statquest]

Thank you! My "to-do" list is about a mile long, so I'll be making videos for many years to come. :)

[@fahdciwan8709]

i'm close to completing my course on Machine Learning, and i owe a ton of gratitude to you. Thanks!

[@statquest]

I'm glad my videos are helpful! :)

[@tymothylim6550]

Thank you, Josh, for this video! I like the Ukulele song at the start, as well as the step-by-step explanation of k-means! It helps a lot that you start with the 1-dimensional case! It makes it much easier to understand!

[@statquest]

Thank you very much! :)

[@adipurnomo5683]

What if I have more than one dimension? Suppose if I have 20 dimension how could I apply that to euclidean distance

[@chaimaa8642]

This youtube channel is a treasure I found, the way he simplifies things .... thank u so much for your amazing explanations

[@statquest]

Glad you think so!

[@Harsh-yw4ct]

You are an absolute legend, this is by far the best video on K-mean clustering on KZbin !!! Thank you so much !!

[@statquest]

Thank you very much! :)

[@jaredcortes366]

This is so wholesome, informative, and engaging all at the same time. Thank you so much for this! Love the intro tune btw

[@statquest]

Thanks so much!

[@Mekkesable]

Your videos are so good, we actually use them in AI Class in University haha. We are doing the reversed Classroom method and the Prof just linked this video for K-Means and we discussed it later. Of course with additional material, but usually he does his own short videos.

[@statquest]

That is awesome!

[@8iter]

You saved me on my final exam for machine learning. Thanks for providing short concise and intuitive videos on ML topics!!! BAM BAM i really like when you say question before saying the question XD

[@statquest]

Thank you very much! :)

[@soumyadeeproy6611]

I really appreciate your efforts in making such good quality video & letting everyone to learn intuitively

[@statquest]

Thank you!

[@Balaji-ci8zw]

When ever iam starting new algorithm, I use to see very easy understanding videos like this after that I will move into deep... Thank you start quest 😊😊

[@statquest]

You're welcome!!! :)

[@saketshetty]

kisine bhi nahi pucha

[@josiahliew5033]

Thank you so much for making these concepts so accessible!

[@statquest]

Glad you like them!

[@anthienaful]

you are a genius, I have never seen more crystal-clear explanations on statistics. THANK YOU

[@statquest]

Hooray!!! I’m glad you like the video! :)

[@kevingaldamez9071]

I never thought I would end up finding a new favorite musician while studying machine learning

[@statquest]

Double BAM! :)

[@aeronwynschache2257]

YOU will the be the SOLE reason I pass my upcoming exam! You have my deep gratitude

[@statquest]

Good luck! Let me know how it goes. :)

[@georgejetson9801]

Listening to your vids before starting a section in the textbook makes learning a happy day! Hooorraaay! for happy brain!

[@statquest]

BAM! :)

[@lllLockonXstratoslll]

the thumbnail wasn't a clickbait, its really clearly explained!. thank you sir

[@statquest]

Thank you! :)

[@liamsnide2057]

I'm just joining the ranks of adoring fans with this comment. Your videos are my first port of call for any new ml and statistics techniques and concepts. You have been a huge help to me.

[@statquest]

THank you! :)

[@henkhbit5748]

Thanks, definitely clearly explained. I would love to see a double BAM on "self organizing map"

[@statquest]

I'll keep that in mind.

[@KeenalShah]

I have been wanting to look at a visual representation for KMeans, so thanks a lot for this!

[@statquest]

You're welcome!!! I'm glad the video was helpful! :)

[@camj92hp]

you explained in 8 minutes what my prof attempted to do in 2 hrs. You are the best!!!!!

[@statquest]

Thanks!

[@lynn1866]

Thanks Josh, the best video to explain statistics concept on KZbin! It is easy for a new starter to understand.

[@statquest]

Glad it was helpful!

[@thongnguyen1292]

StatQuest!!!!!!! StatQuest!!!!!!!!!!!!!! StatQuest!!!!!!!!!!!!!!!!!!!!!! StatQuest!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[@statquest]

That's right!!! :)

[@thongnguyen1292]

Now is an actual comment: Thanks for the great video. Will you cover Gaussian Process in the future?

[@statquest]

Maybe! My to-do list has over 125 things on it so I'm gong to let people vote on what they want me to do next (after I finish up a bunch of machine learning videos).

[@mbv22311]

THIS GOT ME HYPED

[@changfengchen9250]

hooray!

[@ikarus963]

Very clear explanation, perfect ! We would like to listen (and watch :) Mean Shift Clustering from you !

[@krishnaKumar-zi6ct]

Indeed clearly explained in a simple, intiutive manner! Thanks Josh!!

[@statquest]

Thanks! :)

[@roopala5512]

I have my interviews lined up for the upcoming weeks and I have eventually found the best place to prepare for ML related questions. your cute lil introductory song always make me smile :)))) keep making such good stuff

[@statquest]

Best of luck!

[@jahanvin1001]

I was facing anxiety due to my data science project and watching this video, not only I understood it well but that "BAM!!!!!?????" made me crack up so hard lol

[@statquest]

Hooray! :)

[@hkia7893]

Double bam

[@zeetech0123]

Triple bam

[@user-vl6ks6yi9j]

Hi I am just getting into data science but I kinda need someone to mentor me or Just make a fee things clearer that I don’t understand.Do u mind if you could be of help?

[@american-professor]

I use your videos to prepare for my Data Scientist interviews

[@statquest]

Good luck!

[@larryd9577]

Dude. What? This explaination was so thorough and concise and the examples are so well chosen. And perfectly understandably on 2x speed. Well done. And thank you very much!

[@statquest]

Wow, thanks!

[@larryd9577]

​@@statquestNo really. I've seen over the years a lot of tutorials on many different topics and yours achieved more in that 8 minutes than I anticipated. Keep it up. Great format.

[@davidj6755]

This is by far the clearest explanation I’ve seen. Great video!

[@statquest]

Thank you!

[@ryancardenas1873]

amazing. just amazing. please keep up the good work!!! STATQUEST

[@statquest]

Thank you so much! :)

[@ksvrd]

OHHH MY GODDD! Thank youu! Crystal Clear Explanation !

[@statquest]

Hooray! I'm glad you like it. :)

[@patiponwiangnak5698]

You make my research almost complete. Thanks for useful and clearly explanation

[@statquest]

Thanks!

[@read89simo]

these intro's with the uke kinda get you in the mood

[@statquest]

Hooray! :)

[@Pragmaticgeek]

Amazing such a lucid explanation! Thanks a lot & Grateful for whatever you do :)

[@statquest]

Glad you liked it!

[@justinm2015]

my man thats totally dope . i have to agree teaching is obviously an art

[@statquest]

Thank you! :)

[@missregita]

Always love to learn stats with you!! Keep making great videos, Josh! 👏👏

[@statquest]

Hooray!!! I'll keep making videos. It's too much fun. :)

[@missregita]

StatQuest with Josh Starmer I was joining data analysis competition in my college last month and I won the first place! The secret? Playing your videos thru the nights!! That's how grateful I am to you 🙏❤

[@statquest]

No way!!!! That is the coolest thing ever!!! I can't believe it. Wow. I can think of no higher complement than that. Congratulations! :)

[@harleendiaries]

Amazingly explained and I loved that humming in the middle. Such a joyful person, you are. 😌😇 Thank you and all the best.

[@statquest]

Thank you!

[@cw9249]

thank you! but where is the video to explain how to calculate for the "elbow plot" and total variation?

[@corporalwaffles]

"Hoorayyyy!" - I love this part :)

[@statquest]

That's my favorite part too! :)

[@gavincraddock5772]

This is great, I read lots about this but it was far too high level - this is easy to understand,brilliant work. Thanks!

[@statquest]

Glad it was helpful!

[@MMM-wm1vv]

Love your videos!! Could you talk about each algorithm's drawback or limitations? That would be super helpful!!!

[@youzhenzhang8415]

Thank you. I'm in computer vision and I'm absolutely suffering but this at least helped a little! (The most out of everything so far) :)

[@statquest]

Awesome! Good luck with your studies. :)

[@yogeshkumarshankariya642]

Best teacher in the world, love from India.......

[@statquest]

Thanks! :)

[@shubhammishra-ji9oz]

I stumbled upon this video by chance and now StatQuest is the only channel I follow for mathematics and algo for ML.

[@statquest]

BAM! :)

[@alfieheimr]

Gods, three of your videos in this pre-exam morning. Thank you so much. You earned a subscriber.

[@statquest]

Hooray! I hope your exam went well.

[@alfieheimr]

@@statquest It did! An 8 in Data Mining hahaha This afternoon heading towards the final one @o@. Thank you again.

[@faviorc2]

Holy Molly... I would love to learn stats like this.

[@statquest]

:)

[@krisschobert4484]

Another great video Josh! I have one question about the elbow point at 5:51 Is there a formula that determines that the elbow point is at k=3, or does one just have to eyeball it? Maybe one can find the difference in the slopes of the line segments on either side of each kink (each k value). It looks like k = 3 would win at that game. Thus k=3 is the elbow point. Is that what one should do? Thanks again for everything! You're videos are teaching me stats and data science like no one else can!

[@jaredcortes366]

Good point! Would love to know the answer to this too

[@marspark6351]

@@jaredcortes366 also would like to know

[@arwenz1979]

Every time i need you, you're here. you are my hero

[@statquest]

bam!

[@jamesang7861]

as usual, statquest never fails. this would be the only video you need..why did i even bother watching other videos before statquest.

[@statquest]

BAM! :)

[@sahand5277]

..and that is why i turned on notification for this channel. thanks alot!!

[@statquest]

Hooray!!! I'm glad you like the video! :)

[@kanacaredes]

Josh we need a Gaussian Mixture tutorial including EM algorithm!!!

[@statquest]

I'll keep that in mind as well.

[@aryanyekrangi7093]

Dude I love your videos so much, I don't know how I would do my Master's without your videos... *cries*

[@statquest]

I'm glad my videos are helpful! :)

[@iTube4U]

This is by far tbe best intro oftge series, short , focused and realted

[@statquest]

Thanks!

[@michaelcao9483]

"Huh? But what if..?" = "Bam?" "This is amazing stuff" = "BAM!!!" "MINDBLOWING!!!" = "DOUBLE BAM!!!!"

[@statquest]

Thank you! :)

[@amreshwarnaidu7964]

you reminded me of Dwight Schrute. Great video tho, appreciated.

[@statquest]

;)

[@MrKingoverall]

StatQuest: Best machine learning (and stats) tutor PatrickJMT: Best Math tutor

[@Phantom5577]

One of the best and simplest explanations of K-means clustering!!

[@statquest]

Thank you!

[@timto3935]

please do k modes clustering

[@statquest]

I'll put it on the to-do list. Hopefully I can get to it soon. :)

[@ahmedjamel421]

great

[@pH7Programming]

I love the way you explain! Very clear and pleasant to watch 👌

[@statquest]

Thanks so much! 😊

[@Uax101]

I've watched this video a few times in my undergraduate career and the Intro gets me EVERY SINGLE TIME

[@statquest]

BAM! :)

[@naomi7984]

Wow, this was surprisingly simple and fun. Kinda makes me want to marry this guy :)

[@statquest]

Bam! :)

[@arvind31459]

@@statquest Bam! Wam! thank you Mam!

[@dc_8616]

Would be lost without these videos! Thanks.

[@statquest]

Thank you! :)

[@heilyx]

Wow this explanation is flawless. You are one hell of a teacher, thank you :).

[@statquest]

Wow, thanks!

[@SGH-rt6ci]

I like your simple, clear explanation. Thanks a lot.

[@statquest]

Thank you! :)

[@kafikhan8526]

I can't believe how much i love this channel!

[@statquest]

BAM!

[@hariprasadpargunan1]

Great explanation! I would never forget this concept in my life time 😊

[@statquest]

Bam!

[@e-k4110]

StatQuest is awesome, and it comes out every monday

[@statquest]

Almost every monday.... That was a little ambitious!

[@blahb7385]

u sir help me to understand everything i learned for a week into few mins

[@statquest]

Awesome! :)

[@JithuNair95]

I got a campus placement as a Data Scientist and as of yet I've been in the industry for 4.5 years. I am knee deep in everything data science barring ML and AI related intensive coding because I only know the algos at a bird's eye view and my regular work doesn't entail me needing a whole lot of things. But I sat down today to start from K-means and boy am I happy I found this! Such a confidence boost when you understand something really well. I don't even need packages to implement this algo should there ever be a need to do so! Thank you!

[@statquest]

bam! :)

[@tegaobarakpor1029]

This felt like rocket science until today!! thanks!

[@statquest]

bam!

[@user-cv6xl8cb1n]

Thank you very much! I will become a postgraduate student! Your channel is so helpful for me to learn machine learning!非常感谢！！！

[@statquest]

Glad my videos are helpful! :)

[@luistalavera6392]

Your videos are so clear and well explained, thank you so much for this content

[@statquest]

Glad you like them!

[@benjamingwilson]

You have saved my life with this video. THANK YOU.

[@statquest]

Thanks!

[@seungjungjin9217]

Thank you for the clear visuals and explanation!

[@statquest]

Thanks! :)

[@kutta855]

Thanks for visual presentation.. very few videos actually create great visuals. Thanks alot for this!

[@statquest]

You're welcome! I'm glad you like the video! :)

[@highcontrast4488]

Thanks for the explanation, I finally understood this method.

[@statquest]

Glad it helped!

[@arieljiang8198]

I have watched so many of your videos. They are sooooo helpful. My understanding of ML and stats is at a new level now. thank you for making them. You are the best!! could you also make some videos about time series and causal inference methodologies

[@statquest]

I'll keep those topics in mind! :)

[@VioletaHp21]

I've watched those videos from udacity and they explain very well, however yours are the best! :D Will watch them all!

[@statquest]

Thank you! :)

[@mehrabmkg2675]

YOU ARE SAVING ME AT THE NIGHT BEFORE MY EXAM THANK YOU

[@statquest]

Good luck!!! BAM!

[@vcello6450]

You are the BEST stats teacher on KZbin. I will graduate because of you!

[@statquest]

Awesome!!! That's great news. Congratulations! :)

[@hariharanvenkatesan1518]

At the start of the video , Bam? After the end, ohhhhhhh BAM!!! Thanks to you, I am ready for my interview.

[@statquest]

Good luck! BAM! :)

[@hariharanvenkatesan1518]

@@statquest Hey Josh, just to let you know that I aced the interview and got an offer. Double BAM!!. Thanks a lot!!

[@statquest]

@@hariharanvenkatesan1518 Congratulations!!! TRIPLE BAM!!! :)

[@nzerusocia9232]

You, sir, are an awesome teacher. Thank you for making videos as you do.

[@statquest]

Glad you like them!

[@zanaq3]

THANK YOU SO MUCH!! I JUST DID NOT UNDERSTAND THE STEPS BUT NOW IT MAKES SO MUCH SENSE

[@statquest]

bam! :)

[@justsergei5356]

I literally love your explanation. Thank you!

[@statquest]

I'm so glad!

[@jmagames2766]

Amazing!!!!! I was waiting for the double BAAM lol

[@statquest]

Double bam! :)

[@xml6576]

Best tutorial so far !!! Thank you so much~

[@statquest]

Thank you very much! :)

[@flikkie72]

Yay now I can do a Cambridge Analytica! Thanks Josh!

[@statquest]

;)

[@rockyetchison3380]

Loved every min. of video Sir!! Just studied a day before exam & real glad da’t I did 😌

[@statquest]

Good luck!