NOTE: People in different disciplines mean different bases when they write log(x). In Statistics and Machine Learning and most programming languages, the default base is log base 'e'. In other fields, like earthquakes, they use log base 10. Other people use log base 2. So it is important to know what base is being used in the discipline you are working in. Support StatQuest by buying my book The StatQuest Illustrated Guide to Machine Learning or a Study Guide or Merch!!! statquest.org/statquest-store/
Just letting you know that I'm a random german woman in her late 20s who's currently struggling a little with maths and Data Science at Uni but I really enjoy your way of explaining. So glad I found this channel. Wish I had these kinds of videos back when I was studying for my High School grad in 2014, lol...
Honestly.... This is the best video i understood about logs... Till date i wasn't clear where and why do we use logs..... FANTASTIC job... MEGAAAA BAMMMM
never have I seen this much clearest explanation on log. I've seen so many vids but yours is the the THE best...!! Every country should take this video as their log bible.
32+ years of my life and now I know what log does. I am so angry with all my teachers from my school and college. Just yesterday I was asked a question related to log transformation in an interview and I could not answer. I would have answered it had I seen this video. Damn it!
I just wanted to say thank you Josh for your really wonderfully informative and great to understand videos. You have improved the style and learning value of your videos from the early ones to the current ones tremendously. Far less texts and better to follow explanations. They are awesome! You helped me understand more of statistics in a very accessible and instructive way. Btw, I like your style. BAM ON!
Thank you so much! I'm always trying to improve my videos. One day I'd like to go back and re-do all the old ones to make them better. But for now, there is too much new material that I need to cover...
i wanna kiss you on the forehead Josh, my god! PS. Im typing this message from a rural village somewhere in Africa, just know that your videos are making a difference across the world. Im currently learning Data Science so "StatQuest with Josh Starmer" is my bible
Thank you very much for your videos. Everything from basic statistics to machine learning is explained in so nice and easy to understand manner. I have learned here in few days more then during the whole Ph.D.
As an Asian, i (and most people here) learnt logarithm in the first year of highschool, things are pretty simple in school but it has no real-instant application (we mostly practiced with the banking exercises). So we basically forget everything after graduation. Don't know if this also the same with other Western educations.... but anw, thanks for the reminding of the good old knowledge haha - gotta admid your pedagogic methods are just wayyyyyy better than what we have here
Dude. I f***ing love your videos. Thank-you so much for putting these out there and please keep them coming. This is the best explanation I've seen regarding logs; for the first time in my life, I think I understand what logs are and how they are meant to be used. I'm going to share these with my lab. I'm especially excited for your EdgeR part 2 tutorial!
I thought this lesson would be most boring. But as a biochemical engineer, got stonished with your PCR example. Thank you so much. And please come to Sao Paulo next time you come to Brazil!
Such a nice explanation.. first time I have truly understand what logs are and why they are used.. very few people know these fundamentals.. thanks for sharing your knowledge and helping other's to learn and grow in their career.. Really hats off to you and keep doing awesome work.. gbu..
one of the best statquests ever... classic clarity. I would love to see a video explaining when and why natural logs are used.... I recently came across an R package that uses the ln of a ratio... need a quest on it!
In statistics and machine learning, and pretty much every programming language, natural logs are used. The only time we don't use the natural log is when there is some physical reason that another base makes sense. For example, populations of bacteria tend to double over time, so it makes sense to use log base 2 to keep track of those. Perceived loudness (things we hear) is not linear, so we use log base 10 for that. However, when we are just doing math, without some physical basis, we always use log base e, because it makes the math (in general) much, much easier.
FYI: within the Exponential Function,for e to the x, the slope (derivative, tangent line) is equal to itself, e to the x. It's less complicated mathematically than the other exponential forms. Thanks for StatQuest. 👍
Hi! I learned using log when running regression and the data points are all squeezed together. But I never learned the concepts of logs. Tks for this video!
great... help me alot for 4 years after school that now iam working in data processing even young ago i thought it would be useless, now i need them alot 😂
Fold-change is just a figure of speech that means "multiplication". For example, 2-fold larger, means 2-times larger. 2-fold smaller means, 2-times smaller.
My book of logarithmic trig functions are the often negative log+10. For example, to multiply sine times distance, add log.sine(+10 as given) to log.distance then subtract 10 and look up the answer. To divide distance by sine, log.distance + 10 minus log.sine then look up the answer. No negative numbers needed. Both of my logarithm books are base 10. The log of numbers runs from 1 to 100,000 (7 place tables) and can reliably provide 7 significant figures, the 8th is uncertain. The logarithmic sine, cosine, tangents and cotangents are also 7 places, given to 1” of arc. I found these when we moved the office so I’ve been using them, book, pencil and paper.
I have a question ! Let's say I want to see if drinking coffee (vs. not drinking coffee) makes you answer faster to logic related questions. Response time are very small to begin with, so they are very close to 0, and it's therefore common practice to use the log function: log (response time). And I see that there is a difference between drinking coffee and not drinking it. But is it appropriate to make conclusions based on that? What I actually see is that coffee has an effect on the log(response time) not an effect on the actual response time.
Because the log function preserves the order of all the measurements (i.e. if x < y, log(x) < log(y)), if you find a difference between log(x) and log(y), you can conclude that there is a difference between x and y.
Hi Josh, your videos are amazing! Better than a lesson in a classroom. However, I have a few questions about this topic: 1. When should I transform my data into log2? 2. How do my results change if I have raw data vs log2 data? Cheers!
Using log2 depends on the situation. Some data are just naturally meant to have the log transformation and when you plot it, it looks skewed to one side. Using the log transformation on that data makes it look symmetrical. Other times log2 is useful when you want to show 2-fold changes between one set of data an another.
The average of 3.7 you got by dividing (8+2+1)/3 is it? I still could not understand why you skipped the “2 as a baseline generating 4 transcripts” on your third run. Why you jumped from 2 to 8 directly where transcripts are only doubled each run. Secondly how you calculated 1.3 as average in the log land? Please reply.
Yes, 3.7 is the average of 8, 2, and 1. At 8:30 I'm describing 3 different qPCR experiments run for the same number of cycles, not what happens over the course of 3 cycles of PCR. So, the first experiment gives us some amount of transcripts - and we set that has the baseline. We are now interested in the number of transcripts in the next two experiments relative to that baseline. The first of those two gives us twice as many transcripts as the baseline, and the 2nd of those 2 gives us 8 times as many transcripts. Again, these are separate experiments, all run for the same number of cycles.
Hi Josh! Thanks a lot for making these great videos; they are very helpful. I have a question about qPCR (00:08:30): for the third time, why did we ignore 2^2=4; was it not meant to double each time?
Here I'm describing 3 different qPCR experiments run for the same number of cycles, not what happens over the course of 3 cycles of PCR. So, the first experiment gives us some amount of transcripts - and we set that has the baseline. We are now interested in the number of transcripts in the next two experiments relative to that baseline. The first of those two gives us twice as many transcripts as the baseline, and the 2nd of those 2 gives us 8 times as many transcripts. Again, these are separate experiments, all run for the same number of cycles.
Thank a lot Josh!, I am miguel pastor from Perú. I have a questions. When a I am doing a regression with many outliers in the dependet variable, this transformation to Log scale will help to hace better results(P Value, ecuation of predicction, etc)?
I don't know. If they really are outliers, then regardless of the transformation, you should probably remove them. Usually, people transform their data with the log if it makes sense in a general way - for example, if taking the log of the data results in it being normally distributed.
I have a question (please correct me if I misunderstood) !! @8:30 - The third time we do qPCR, it would again double (from that in the second time) and hence shouldn't it be as many transcripts as the first time ? 1st time -> 2^0 = 1, second time -> 2^1 = 2 and hence the third time 2^2 = 4.
I think you misunderstood the example. These are just replicates of the same experiment, where we run the machine the same number of cycles each time. However, each time we do the experiment, we get a different result because we started with different amounts of RNA.
@@astrovishalthakur The example is not cycles, but different samples. Each time we do the experiment, we do the exact same number of cycles. However, each sample starts with different amounts of RNA that we are amplifying. The 3rd sample starts with 8 times as much RNA as the 1 sample, so, after cycling the machine (possibly 10 times, for each sample) we end up with enough RNA that we can detect the 8 fold difference.
I have a question if were checking to see if data is normally distributed we may do a linear transformation to normalize the data. Should we first remove outliers through other methods before applying a transformation?
Thank you so much! now I know logFC = 0 is actually a positive FC 1. I have something still unclear: say if I have a RNAseq dataset with tons of 0s/low values as well as some extremely high values, then when I use log transformation, which base should I choose? I have seen other people use log2 heavily for microarray data, but why is that? Big thank you!
I think there is a typo at about 8:32. I think it should be the fourth time instead of the third time because the first time is 1=2^0 instead of 2=2^1.
The video is correct - by "time", I'm not talking about PCR cycles, I'm talking about different PCR experiments. We did 3 separate PCR experiments, and each experiment yielded different results.
i watching several your video.these videos is very best.thanks a lot of.I do not know just how to watch videos. Which movie should I start from? My education is my statistics
Could you recommend some pre subject matter that is important in order to understand logarithms? I've been trying to grasp the concept for years but I can't understand it.
I'm not sure what you are asking about. However, we said that the third time there were 8 times as many transcripts. 2^3 = 8, so the log2(8) = 3. Does that make sense?
Hey! Can you put the ads at the beginning? It's frustrating setting it down and pulling out my pencil and then having to go skip. Just set it at the start like everyone else.
In this example, I say "run" to mean the entire qPCR experiment. As in "we ran the entire qPCR program". You may be confusing my use of the word "run" with "cycle" - transcripts double each cycle. The first experiment is the baseline, so we call that '0'. The second experiment had twice as many transcripts as the first experiment, so that's 2^1 on the normal scale, and 1 on the log scale - this means that there is a 1 cycle difference between the 1st and 2nd experiment. The third time we do the experiment, there are 8 times as many transcripts. This is 8 on the normal scale and 3 (since 2^3=8) on the log scale. This means that there is a 3 cycle difference between experiments 1 and 3. Does that make sense?
@@statquest Josh, why do we count cycles difference? I thought the number of cycles in each test identical. And the goal is - to distinguish original number of transcripts after applying identical number of amplifications?
@@SergeySenigov It's just terminology. How many cycles do we need to run to end up with the exact same number of transcripts? It's that difference in cycles that we report.
![ทุกครั้งที่หลับตา (Lucid Dream) - AYLA's [ Official MV ]](http://i.ytimg.com/vi/4i1OBAQyv58/mqdefault.jpg)
NOTE: People in different disciplines mean different bases when they write log(x). In Statistics and Machine Learning and most programming languages, the default base is log base 'e'. In other fields, like earthquakes, they use log base 10. Other people use log base 2. So it is important to know what base is being used in the discipline you are working in.
Support StatQuest by buying my book The StatQuest Illustrated Guide to Machine Learning or a Study Guide or Merch!!! statquest.org/statquest-store/
I ve tried to buy ur book. But local bank card not working. May b paypal will work. Please tell me wat shud I buy as m a data science aspirant
@@vaibhavpandey7398 PayPal should work.
Seriously, If I cry today, please don't stop me. God bless you, Josh.
Thanks! And thank you for your support!
Just letting you know that I'm a random german woman in her late 20s who's currently struggling a little with maths and Data Science at Uni but I really enjoy your way of explaining. So glad I found this channel. Wish I had these kinds of videos back when I was studying for my High School grad in 2014, lol...
Thanks! I'm glad you enjoy my videos. BAM! :)
12yrs to get this because every explanation until now has been shoddy at best, or deliberately obtuse at worse. Thank you
Hooray! I'm glad this video was helpful.
Honestly.... This is the best video i understood about logs... Till date i wasn't clear where and why do we use logs..... FANTASTIC job...
MEGAAAA BAMMMM
never have I seen this much clearest explanation on log. I've seen so many vids but yours is the the THE best...!! Every country should take this video as their log bible.
:)
32+ years of my life and now I know what log does. I am so angry with all my teachers from my school and college. Just yesterday I was asked a question related to log transformation in an interview and I could not answer. I would have answered it had I seen this video. Damn it!
Hopefully your next interview will go better. :)
I just wanted to say thank you Josh for your really wonderfully informative and great to understand videos. You have improved the style and learning value of your videos from the early ones to the current ones tremendously. Far less texts and better to follow explanations. They are awesome! You helped me understand more of statistics in a very accessible and instructive way. Btw, I like your style. BAM ON!
Thank you so much! I'm always trying to improve my videos. One day I'd like to go back and re-do all the old ones to make them better. But for now, there is too much new material that I need to cover...
I really love you and all your videos. I've learned more in a day than what I've learned in a whole month in the lab.
Hooray!!!! I'm glad you like the videos! :)
i wanna kiss you on the forehead Josh, my god!
PS. Im typing this message from a rural village somewhere in Africa, just know that your videos are making a difference across the world. Im currently learning Data Science so "StatQuest with Josh Starmer" is my bible
BAM!!! That's awesome!!! Good luck with your studies!
Thank you very much for your videos. Everything from basic statistics to machine learning is explained in so nice and easy to understand manner. I have learned here in few days more then during the whole Ph.D.
Thank you very much! :)
As an Asian, i (and most people here) learnt logarithm in the first year of highschool, things are pretty simple in school but it has no real-instant application (we mostly practiced with the banking exercises). So we basically forget everything after graduation. Don't know if this also the same with other Western educations.... but anw, thanks for the reminding of the good old knowledge haha - gotta admid your pedagogic methods are just wayyyyyy better than what we have here
Thank you very much! I'm glad my video was helpful.
After so many years, this video still delivers!!!!
bam! :)
I just unintentionally got the best interpretation of geometric mean. Super helpful video. Thanks
bam!
You don't even know how much you are helping me survive grad stats 3. Thank you, thank you, thank you. I subscribed.
This video helped me so much in my maths IA! Not all heroes wear capes, and you're living proof 🙌 ¡¡¡¡muchas gracias!!!!
De nada!! :)
Beautifully explained. For the first time in my life, I was able to understand why logarithm is used.
bam! :)
Dude. I f***ing love your videos. Thank-you so much for putting these out there and please keep them coming. This is the best explanation I've seen regarding logs; for the first time in my life, I think I understand what logs are and how they are meant to be used. I'm going to share these with my lab.
I'm especially excited for your EdgeR part 2 tutorial!
I thought this lesson would be most boring. But as a biochemical engineer, got stonished with your PCR example. Thank you so much. And please come to Sao Paulo next time you come to Brazil!
Awesome!!! I'm dreaming of going to Sao Paulo. I will go there next time! :)
The best statistical tutorial on TH-cam! BAM!!
Such a nice explanation.. first time I have truly understand what logs are and why they are used.. very few people know these fundamentals.. thanks for sharing your knowledge and helping other's to learn and grow in their career..
Really hats off to you and keep doing awesome work.. gbu..
Thank you!
This is so simple yet so brilliant.
Wish I had a teacher in school who taught us such stuff.
Thanks!
You deserve a Nobel Prize in Acedemics.😎
bam!
Damn, thank you so much!! This was very, very useful as a person who always remembered the rules but did not understand them.
Bam! :)
8th wonder in the history, I understood log, hurray! I love you bro
bam!
Great lesson! So easy to grasp and with practical examples.
Thank you! :)
one of the best statquests ever... classic clarity. I would love to see a video explaining when and why natural logs are used.... I recently came across an R package that uses the ln of a ratio... need a quest on it!
In statistics and machine learning, and pretty much every programming language, natural logs are used. The only time we don't use the natural log is when there is some physical reason that another base makes sense. For example, populations of bacteria tend to double over time, so it makes sense to use log base 2 to keep track of those. Perceived loudness (things we hear) is not linear, so we use log base 10 for that. However, when we are just doing math, without some physical basis, we always use log base e, because it makes the math (in general) much, much easier.
@@statquest Thanks Josh... I think when i start doing a few problems with it, i will start to grasp it better.... thats usually what happens!
FYI: within the Exponential Function,for e to the x, the slope (derivative, tangent line) is equal to itself, e to the x.
It's less complicated mathematically than the other exponential forms.
Thanks for StatQuest.
👍
This is the best explanation of logarithm I ever saw. And of course your singing make a relaxed atmosphere :) Thank you very much!
What an explanation man. This is gold. Thank you so much :)
Thank you! :)
Hey Josh I started to learn data science things. your lectures are super duper useful for me to understand. Thx a lot
Happy to help!
10 seconds into your explanation of logs and everything clicked. Wow
bam! :)
It is been 16 years I have studied Math, now for DS I have started again. But LoL today I knew why we have Log. 16 BAMs to you Sir.
Hooray! :)
Extremely well made vid, Keep the math magic flowing! ✨
Thank you! Will do!
Thank you so much. You're a wonderful educator. You're helping me do my job 😅
Happy to help!
Awesome you are Josh Starmer. Helping you did. Logs I understand. Amazing is your singing.
BAM! :)
Fast forward to 2021... and the PCR test is front and center in the world of Covid. Thanks for the video!
Bam!
You guys have a video for everything. And you make me laugh! Thank you for this amazing channel, I'll pass my exams because of you!
Good luck on your exams!
Best explanation about log🔥
Thank you!
Hi! I learned using log when running regression and the data points are all squeezed together. But I never learned the concepts of logs. Tks for this video!
great... help me alot for 4 years after school that now iam working in data processing even young ago i thought it would be useless, now i need them alot 😂
:)
Big BAM is not big enough for this content.
Thanks! :)
Map BAM on log scale to the base 100
Fantastic videos I wish I had this video when teachers were asking me to mug things up. Can you clarify what is fold change?
Fold-change is just a figure of speech that means "multiplication". For example, 2-fold larger, means 2-times larger. 2-fold smaller means, 2-times smaller.
RT-PCR: who would have thought that something asked to imagine by Josh would become reality in 2020!
BAM! :)
that has to be the best statquest song
it's a good one! :)
all video are great to understand, please come up with other concepts as well
Me, watching this video at 3:25 AM:
The video intro:
Me: "Same"
Bam! :)
the king of intuition
Thank you! :)
This video was great, thanks Josh!
Bam! :)
I subscribed just because of the intro music
You have excellent taste in music! ;)
“In theory it’s natural number, but I always forget why” I was math major but this is so true😂
Yes! :)
This is `e`. :) --> th-cam.com/video/_-x90wGBD8U/w-d-xo.html
My book of logarithmic trig functions are the often negative log+10. For example, to multiply sine times distance, add log.sine(+10 as given) to log.distance then subtract 10 and look up the answer. To divide distance by sine, log.distance + 10 minus log.sine then look up the answer. No negative numbers needed.
Both of my logarithm books are base 10. The log of numbers runs from 1 to 100,000 (7 place tables) and can reliably provide 7 significant figures, the 8th is uncertain. The logarithmic sine, cosine, tangents and cotangents are also 7 places, given to 1” of arc. I found these when we moved the office so I’ve been using them, book, pencil and paper.
noted!
When the solar EMP hits I’m going to be ready.
@@someonespadre TOTES!
Btw I shared this channel on my college piazza forum and people are going crazy :)
Bam! Thank you very much. Sharing is the best complement ever! :)
Hi Josh, thank you very much for the insightful video. I was wondering whether you could add one on Lognormal distribution? :) Thank you!
I'll keep that in mind.
I have a question !
Let's say I want to see if drinking coffee (vs. not drinking coffee) makes you answer faster to logic related questions.
Response time are very small to begin with, so they are very close to 0, and it's therefore common practice to use the log function: log (response time).
And I see that there is a difference between drinking coffee and not drinking it. But is it appropriate to make conclusions based on that? What I actually see is that coffee has an effect on the log(response time) not an effect on the actual response time.
Because the log function preserves the order of all the measurements (i.e. if x < y, log(x) < log(y)), if you find a difference between log(x) and log(y), you can conclude that there is a difference between x and y.
omg thank you! I also bought your book like I mentioned before :)
Thank you! :)
Really it was a very exciting video.
There's none identical. It's unique
you ! are! MY HERO
Thanks!
u hu hu BEST u hu hu INTRODUCTION u hu hu EVER
Yes, it's a good one! :)
Hi Josh, your videos are amazing! Better than a lesson in a classroom. However, I have a few questions about this topic: 1. When should I transform my data into log2? 2. How do my results change if I have raw data vs log2 data? Cheers!
Using log2 depends on the situation. Some data are just naturally meant to have the log transformation and when you plot it, it looks skewed to one side. Using the log transformation on that data makes it look symmetrical. Other times log2 is useful when you want to show 2-fold changes between one set of data an another.
The average of 3.7 you got by dividing (8+2+1)/3 is it? I still could not understand why you skipped the “2 as a baseline generating 4 transcripts” on your third run. Why you jumped from 2 to 8 directly where transcripts are only doubled each run. Secondly how you calculated 1.3 as average in the log land? Please reply.
Yes, 3.7 is the average of 8, 2, and 1.
At 8:30 I'm describing 3 different qPCR experiments run for the same number of cycles, not what happens over the course of 3 cycles of PCR. So, the first experiment gives us some amount of transcripts - and we set that has the baseline. We are now interested in the number of transcripts in the next two experiments relative to that baseline. The first of those two gives us twice as many transcripts as the baseline, and the 2nd of those 2 gives us 8 times as many transcripts. Again, these are separate experiments, all run for the same number of cycles.
Hi Josh! Thanks a lot for making these great videos; they are very helpful. I have a question about qPCR (00:08:30): for the third time, why did we ignore 2^2=4; was it not meant to double each time?
Here I'm describing 3 different qPCR experiments run for the same number of cycles, not what happens over the course of 3 cycles of PCR. So, the first experiment gives us some amount of transcripts - and we set that has the baseline. We are now interested in the number of transcripts in the next two experiments relative to that baseline. The first of those two gives us twice as many transcripts as the baseline, and the 2nd of those 2 gives us 8 times as many transcripts. Again, these are separate experiments, all run for the same number of cycles.
@@statquest Got it, thanks!
As usual awsome explanations.
Thank you! :)
very good explanation
Thanks! :)
Wooow mind blowing sir. Thank you so much for this video .
Thanks!
Thank a lot Josh!, I am miguel pastor from Perú. I have a questions. When a I am doing a regression with many outliers in the dependet variable, this transformation to Log scale will help to hace better results(P Value, ecuation of predicction, etc)?
I don't know. If they really are outliers, then regardless of the transformation, you should probably remove them. Usually, people transform their data with the log if it makes sense in a general way - for example, if taking the log of the data results in it being normally distributed.
I have a question (please correct me if I misunderstood) !!
@8:30 - The third time we do qPCR, it would again double (from that in the second time) and hence shouldn't it be as many transcripts as the first time ?
1st time -> 2^0 = 1,
second time -> 2^1 = 2
and hence the third time 2^2 = 4.
I think you misunderstood the example. These are just replicates of the same experiment, where we run the machine the same number of cycles each time. However, each time we do the experiment, we get a different result because we started with different amounts of RNA.
@@statquest It's still confusing, can you explain again how on 3rd cycle it reached 8 but not 4.PLEASE...
@@astrovishalthakur The example is not cycles, but different samples. Each time we do the experiment, we do the exact same number of cycles. However, each sample starts with different amounts of RNA that we are amplifying. The 3rd sample starts with 8 times as much RNA as the 1 sample, so, after cycling the machine (possibly 10 times, for each sample) we end up with enough RNA that we can detect the 8 fold difference.
Great video
Thanks!
Thnx sir for your valuable information
Thanks!
feynman would appreciate this channel
BAM! :)
I have a question if were checking to see if data is normally distributed we may do a linear transformation to normalize the data. Should we first remove outliers through other methods before applying a transformation?
If you can identify outliers, you should always remove them as early as possible.
For a moment i thought..coldplay started teaching stat or what!
bam!
Thank you so much! now I know logFC = 0 is actually a positive FC 1. I have something still unclear: say if I have a RNAseq dataset with tons of 0s/low values as well as some extremely high values, then when I use log transformation, which base should I choose? I have seen other people use log2 heavily for microarray data, but why is that? Big thank you!
I have weekly sales data which contains some negatives and zero values. What is the best way to handle them in your opinion?
I don't know.
Natural Log, e stands for exponential relationship as in calculating compound interest
nice
Hi , Thanks for your videos, these are really helpful. Can you make one video on vector and matrix and where we use them in practical approach.
One day I would like to do that.
thank you for the great work !
Thanks!
keep the good work
Thanks! :)
Thank you for this. Really good.
Thanks! :)
god bless this man
Thanks!
Thanks a lot!
Thanks!
Beautiful.
Thanks!
I think there is a typo at about 8:32. I think it should be the fourth time instead of the third time because the first time is 1=2^0 instead of 2=2^1.
The video is correct - by "time", I'm not talking about PCR cycles, I'm talking about different PCR experiments. We did 3 separate PCR experiments, and each experiment yielded different results.
@@statquest I see, thanks for your clarification.
i got them all except the point related to the mean of logs i hope you can use different example , thank you for the great explain
I'll keep that in mind.
A better choice of song lyrics would be "Stat Quest Stat Quest whatcha gonna do? Whatcha gonna do to compute log(u)" 😁
:)
i watching several your video.these videos is very best.thanks a lot of.I do not know just how to watch videos. Which movie should I start from?
My education is my statistics
i learnt it at 17 yrs but i get it at 33 with this video.
Awesome! :)
Could you recommend some pre subject matter that is important in order to understand logarithms? I've been trying to grasp the concept for years but I can't understand it.
Can you tell me what time point (minutes and seconds) in this video where you got confused?
Great videos
Thanks!
Remember e from math class? I don't, me neither hahahah you are amazing
:)
Fantastic!!!!!!
Thank you! :)
very nice!
Thanks!
5:43 why did we measure the distance of 8 and 1/8 from 1 but the exponent difference from 0?
Because the log(1) = 0.
I have a question that should I log transformate my mass spectrum data for principal component analysis?
❤Thank you~
The third when we do the qpcr then the no of transcripts should be 4 right ?
I'm not sure what you are asking about. However, we said that the third time there were 8 times as many transcripts. 2^3 = 8, so the log2(8) = 3. Does that make sense?
@@statquest Got it thanks
Hey! Can you put the ads at the beginning? It's frustrating setting it down and pulling out my pencil and then having to go skip. Just set it at the start like everyone else.
I'll keep that in mind.
Thank you!
:)
Shouldn't the third time we do the qPCR we should have 4 times as many transcripts as the first time?
In this example, I say "run" to mean the entire qPCR experiment. As in "we ran the entire qPCR program". You may be confusing my use of the word "run" with "cycle" - transcripts double each cycle. The first experiment is the baseline, so we call that '0'. The second experiment had twice as many transcripts as the first experiment, so that's 2^1 on the normal scale, and 1 on the log scale - this means that there is a 1 cycle difference between the 1st and 2nd experiment. The third time we do the experiment, there are 8 times as many transcripts. This is 8 on the normal scale and 3 (since 2^3=8) on the log scale. This means that there is a 3 cycle difference between experiments 1 and 3. Does that make sense?
@@statquest Josh, why do we count cycles difference? I thought the number of cycles in each test identical. And the goal is - to distinguish original number of transcripts after applying identical number of amplifications?
@@SergeySenigov It's just terminology. How many cycles do we need to run to end up with the exact same number of transcripts? It's that difference in cycles that we report.
Fantastic
Thanks!
Is there a book you can suggest which explains multivariate statistics as lucidly as your videos.....
Unfortunately I can't.