Some extra things that would be worth mentioning: 1) in the absence of linearity, the correlation itself may give you very little information. In some cases, data set with Corr = 0.9 might be closer to the one with Corr = 0, than to the other with Corr = 1. In other words, very weak trends (maybe not even worth mentioning) may give you high Corr value. 2) In random generated sets, certain degree of correlation may appear randomly, without a good reason. 3) Not only does correlation not imply causation, causation doesn't imply correlation as well. Take for example y = |x| for x \in [-1, 1]. Clearly y=f(x), yet Corr = 0.
This is literally the greatest explanation that I've ever heard. Beautifully animated and very straight to the point. The best video ever on this topic! You saved my life!😊
I’m a statistics student in my final year of undergrad. I just felt like checking out a video on covariationnas we begin PCA. This is by far the most clear it has ever been explained, and so succinctly! Well done, I feel I can think about it more clearly now.
RUclips can't be serious that this video doesn't top their search result! Kudos to the channel. This video seriously deserves every compliment this world can offer. Keep up the great work thanks!
6:346:34 For example, let's say we observe a correlation between the number of ice cream cones sold at a beach and the number of drownings. There might indeed be a correlation, as both tend to increase during hot summer days. However, it would be incorrect to conclude that eating ice cream causes drownings or vice versa. Instead, there could be a third variable, such as temperature, that influences both variables independently.
Fantastic explanation, but one of the things that really sets this apart from some of the others I have seen is that you take time at the end to show the different values for different points. Really a great way to both help people visualize and check if they understood correctly.
Amazing intuition and excellent animation! Interestingly, the sign of correlation follows the monotony of sine function in each of the 4 quarters of the unit circle. In the top right quarter [0, 90deg] the sine function is monotonically increasing and correlation is positive, in the top left quarter [90, 180deg] the sine function is monotonically decreasing and correlation is negative and so on.
@@NormalizedNerd Ah! Actually not. In the bottom two the behavior is reversed! The sign of tangent in each quadrant would have been a better example. :)
that is exactly what i've been looking for for along time ,,, very clear and a big thumb for everything but espessially for the last visualization .....
The video looks more like an animation exercise rather than an educational video. The formulas are explained but the intuition behind them is as opaque as before. A few videos from 3Blue1Brown could be a good reference for this channel.
1:37 Hello sir can I know how you drew that type of Graph with Manim? I just want the name or line of the code about how you drew that kind of grid please sir!
The fact that I (a math teacher) learnt something new from this, means my students need to see it. Thumbs up my friend
You just made my day! ❤
Some extra things that would be worth mentioning:
1) in the absence of linearity, the correlation itself may give you very little information. In some cases, data set with Corr = 0.9 might be closer to the one with Corr = 0, than to the other with Corr = 1. In other words, very weak trends (maybe not even worth mentioning) may give you high Corr value.
2) In random generated sets, certain degree of correlation may appear randomly, without a good reason.
3) Not only does correlation not imply causation, causation doesn't imply correlation as well. Take for example y = |x| for x \in [-1, 1]. Clearly y=f(x), yet Corr = 0.
Great comment! May be I will make a dedicated video on Correlation in the future :D
This is literally the greatest explanation that I've ever heard. Beautifully animated and very straight to the point. The best video ever on this topic! You saved my life!😊
I’m a statistics student in my final year of undergrad. I just felt like checking out a video on covariationnas we begin PCA. This is by far the most clear it has ever been explained, and so succinctly! Well done, I feel I can think about it more clearly now.
Could you explain why eigenvectors corresponding to largest eigenvalues maximize projected variance intuitively?
really asking
RUclips can't be serious that this video doesn't top their search result! Kudos to the channel. This video seriously deserves every compliment this world can offer. Keep up the great work thanks!
The best explanations and visualizations I've ever seen. Thank you for that!
6:34 6:34
For example, let's say we observe a correlation between the number of ice cream cones sold at a beach and the number of drownings. There might indeed be a correlation, as both tend to increase during hot summer days. However, it would be incorrect to conclude that eating ice cream causes drownings or vice versa. Instead, there could be a third variable, such as temperature, that influences both variables independently.
The animations helped me get an idea of what I was calculating for this long :) Maths is beautiful✨
So clear. Thanks for explaining. Also the background music was super calming
Fantastic explanation, but one of the things that really sets this apart from some of the others I have seen is that you take time at the end to show the different values for different points. Really a great way to both help people visualize and check if they understood correctly.
Excellent video. Very intuitive and well explained! As someone who's both a student and a teacher of maths, I'm impressed from both angles!
I agree; this video was amazing! And you are so right- the video, in the end, was so good to see what the values look like at different plot points.
great video! you really helped me to understand in 7 minutes what my proffesor could not in 1 hour! thanks!
Actually such a good video, the visual representations helped me finally get the inuition i was missing for covariance. Thank you!
what a beautiful animation, amazing video friend!
Instantly subscribed. Please keep up the work! Amazing visualization.
this is the kind of explanation we need . 🙂
Incredible explanation and visualization! Thank you!
Amazing intuition and excellent animation! Interestingly, the sign of correlation follows the monotony of sine function in each of the 4 quarters of the unit circle. In the top right quarter [0, 90deg] the sine function is monotonically increasing and correlation is positive, in the top left quarter [90, 180deg] the sine function is monotonically decreasing and correlation is negative and so on.
Thanks mate!
Is this observation true for bottom two quadrants?
@@NormalizedNerd Ah! Actually not. In the bottom two the behavior is reversed! The sign of tangent in each quadrant would have been a better example. :)
It's such a great explanation! Thank you so much!
This is so well explained. Thank you so much
the visual is sick!! amazing
that is exactly what i've been looking for for along time ,,, very clear and a big thumb for everything but espessially for the last visualization .....
The video looks more like an animation exercise rather than an educational video. The formulas are explained but the intuition behind them is as opaque as before. A few videos from 3Blue1Brown could be a good reference for this channel.
Great video, thank you! I finally understand covariance. Very helpful and beautiful visualizations as I study for my statistics final :)
great video ! i really appreciate the effort you put into this
Great video. You really did explain the concept in a very elegant way.
Thanks for this video, you really helped me understand covariance visually.
thank you sir, great teaching and visualizations. 👍😊
First time I comment a video of a math explanation in years, wow very good!
amazing video, so happy I came across it
Great video , thanks!!
super good animaitions and music
thank you clearly explained.
Thsnk you much. Please more statistics.
Can't wait for your channel to get discovered
Well explained.
Can you tell me why we perform eigen decomposition to the covariance matrix in PCA ?
Good animation, well explained, thanks for sharing 👍
well explained !
1:37 Hello sir can I know how you drew that type of Graph with Manim? I just want the name or line of the code about how you drew that kind of grid please sir!
For creating the grid you need to use NumberPlane object.
@@NormalizedNerd Thanks man
Excellent explanations!
thanku so much man
Very helpful thank you!
thank you it is clear!
hello, anyone can explain what is the yellow dot lines ? what does it represent?
Damm next level teaching
your definition about covariance is partially correct covariance also shows the strength of the relationship but not the linear relationship
very good video!
It's crazy learning correlation without telling you about covariance
Very intuitive, thanks!
This video is well explained than my teacher
Brilliant animations, thanks
Thanks mate!!
This is art
great video!
eloquently explained
well explained
amazing video
Thank you
amazing and satisfying
Brilliant!
Thanks!
Best explanation..👍..sir, please make videos on regression line.....🙏
Thanks!!
this is the best thin ever!
Do you use manim?
Beautiful
Please do degrees of freedom
It’s truly intuitive … is there a way I can connect with you?
I recognise a lot of 3Blue1Brown's animation and audio style here ^^ nevertheless, thanks for the helpful video
Yeah, I'm using his python library!
That is beautiful 😢
holy shit that was incredible
thankyou
Nice animation
tyvm
Amazing
super video
thanks!!!
Perfect!
Thank you..
Limpid. Great job.
Thanks a lot :D
Lovely animations, 3B1B style c:
Thanks a lot! Yup using his library.
Eventually somebody did it ! :)
wow