Top 5 Statistics Concepts in Data Science Interviews: P-value, Confidence Interval, Power, Errors

[@songxiyou2347]

自己复习才发现，Emma真是将这些内容完全吃透，整理成自己的体系。不管是product sense还是stat，全部是干货并且非常organized。多余的废话一句没有（对比我自己的录音回答发现了一堆废话hhh）。非常感谢行业内有这样的领路人。继续期待product sense实例分析/stat & probablity 考点/take home & presentation思路总结和其他DS相关内容！Emma 新年快乐！新的一年身体健康，工作顺利，万事如意！

[@徐宁远]

Super useful. One of the best DS videos I have ever seen !

[@katekatebangbang2435]

作为一个在面试的人，来回来去看了好多次emma的视频了，常看常新。谢谢Emma

[@zihenglin5294]

Thought I already known those stats concepts but still learned a lot from your video. The tips for technical and non-technical audience are very helpful! Thanks Emma. Love your content!

[@anathemaconscience5666]

hi emma, i am kind of confused to the p value. At 10:33 you mentioned small p, more convinced of difference. But at 11:22, you said p value represents there is a diff given null hypo is true, meaning higher p, more convinced of difference. But given the height example, i believe small p larger difference, so at 11:22, why would you say p means there is a diff given null hypo is true?

[@yuanliu2496]

I came across your video and it turns out to be super helpful! Thank you! subscribed.

[@mussdroid]

We are going to moon on Data Science 🚀🚀🚀🚀 🌜🌜🌜 ! Thanks Emma

[@sitongchen6688]

This is super clear, and now I have a good sense or expectation from the interviewer! Thanks Emma!

[@josephjoestar995]

So glad I came across this goldmine of a channel, honestly such great relevant topics with the most useful explanations - I trust you 100% to help with my interviews haha

[@fengzhoupan771]

Love the video! Thank you so much for the tips!

[@insigh01]

The way you structure your response is concise, and it makes it easy to understand these concepts. Thank you Emma!

[@yenliknurasheva6322]

I am very grateful for your useful videos! Great content! You are so smart and beautiful! 😇 Also preparing for DS interview, these videos help a lot!!!

[@281019641]

Thanks Emma. Very clear description and helpful to see the categorization accordingly for technical and non-technical audience.

[@Nancy-wr7zb]

Great video Emma !! Technical vs non technical explanations were very impressive !!

[@spotting_experiment]

Landed here preparing for my upcoming interview and this is very useful as a revision material as well.

[@jayzune1752]

Wooo, smart and elegant lady! Thanks for your video, helped me a lot!

[@hehuang3536]

Hi Emma, I have watched a lot of videos you made and they are super clear and helpful for preparing my DS interviews. Thank you so much!

[@emma_ding]

Hey, I'm so happy to hear that my videos have been helpful. Best of luck with your interviews!

[@goodjuju2132]

Emma thank you so much for all of your quality content!! You're doing so much for the community

[@jaden2582]

NO one word of bullshit. Appreciate it, Emma.

[@jingyou3481]

This is really great. I've been thinking about how to explain p value to non-technical person and find a great example for a while. This is definitely very clear! Hope you can continue to make some videos for stats concept like Simpson Paradox etc

[@yinqiu6780]

So well explained! Thank you Emma!

[@weiyang2116]

Yay! Exactly what I was looking for! Thanks Emma

[@nisithaukkarapattanakul8860]

Very clear explanation, thanks

[@ishitasadhukhan1]

Amazing videos Emma ! I am preparing for data science interviews and feel so lucky and grateful that I found your channel ! I am making it a point to follow your advice to the words ! Thank you so much for what you are sharing with us!

[@DataProfessor]

Thanks Emma! Awesome video also for practicing data scientists, it’s a great video to brush up on our stats knowledge 😆

[@emma_ding]

Thank you Data Professor!

[@niveditakumari701]

Thank you for the video, can you please share another example for p-value in the layman's term?

[@Mackymon]

Great Vid! Follow up question: how do you get a feel for how technical your audience actually is?

[@emma_ding]

Look at their public profile like LinkedIn :)

[@hehuang3536]

In one of my technical interviews, the interviewer asked me how do you explain the concept to your grandma?

[@mihirbosemj]

The content you publish is so helpful for us to learn data science and prepare for interviews. Keep up the great work, and all the best :-)

[@alifiaz7792]

Very intuitive video. Please also consider making a video explaining the metrics for regression, classification and clustering machine learning models from both technical and business perspective.

[@Sethsm1]

Extremely helpful. Thank you.

[@wongkitlongmarcus9310]

thank you Emma

[@muse3324]

1:41 "It should not be obscure like what you see in Wikipedia" 😅😁😁

[@abhi_and_seki]

Very informative and helpful ❤

[@emma_ding]

So happy to be of assistance, Mrinal! 😊

[@abhi_and_seki]

@@emma_ding just ended up with my data scientist internship interview and it was very very good. Thankyou for such amazing content. It was very helpful for last minute brushup of key skills and i am hoping for positive results from my interviewer 🤞✨

[@emma_ding]

That's fantastic to hear, Mrinal! Feel free to keep me posted with how your results go. Fingers crossed, and sending you good luck! 💛

[@abhi_and_seki]

​@@emma_dingThankyou so much for the good wishes and all your hard work in videos was worth it because we benefited from them a lot. Also, I would like to share that I have accepted the Data Scientist Internship with Loblaws Companies in Toronto, Canada, for the coming Winter of 2023. I am so excited and obliged to start my new journey in Data Science. It was difficult but with consistent hard work and good resources such as your channel, I am now going to follow my dream career. Thankyou once again for all good work and keep posting such insights and helpful resources on DS, as it will still help me during my professional career.

[@emma_ding]

Mrinal! This is fantastic news! Thank you for sharing this huge win with me, and congratulations on your new role. I can't wait to hear what else is in store for you in the future. Sending you all the best! 🥳

[@hameddadgour]

Great content!

[@taozhang7696]

thank you. it's really helpful!

[@jiayiwu4101]

Wow, super cool summary! Really practical! Thanks Emma. Would you mind sharing slides or text then?

[@emma_ding]

Sorry there are no slides. It's part of the video editing.

[@bhageerathbogi4951]

Hi Emma, Can you please share a link to the slides.

[@emma_ding]

Sorry, there's no slides, it's all part of the video editing. But I'll definitely consider providing it in the future if it helps!

[@jennywu799]

Emma, 可以不可以出一个视频总结一下常用的distribution，有的时候面试的时候被问到sales data是什么样的distribution，我每次都答normal。。。

[@jaden2582]

poisson distribution

[@yogiHalim]

Significance (p-value 80%) is the probability of correctly [rejecting the null hypothesis while it is false.]. (probability of not testing positive pregnancy for male) for 3 or more outcome, [testing negative] >< [not testing positive]. Significance is thus the probability of Type I error, whereas 1−power is the probability of Type II error.

[@aliciama1745]

really helpful! Thank you very much for do this! Emma, can you introduce * how to do a project* for the people who want to transfer to data science from other unrelated fields? Appreciate ahead of time!

[@emma_ding]

For learning purpose, Kaggle is a really place to start. For "real-life" projects, you have to look for opportunities of side projects or in your current position.

[@liumx31]

Hi Emma, thanks for the great explanation, one question though -- how is power used to determine the sample size? I thought the sample size determined the power, i.e. the larger the sample size the higher the statistical power.

[@mattpras]

great video!

[@guimaraesalysson]

Great video, helps a lot

[@jeoffleonora4612]

Well explained. Thank you!

[@plttji2615]

What if N increase, does it affect P-value?

[@LouisChiaki]

10:09 some typo on the slides. Should be "null" not "hull" hypothesis :D

[@emma_ding]

Thanks for catching the typos!

[@chengqian5737]

给你一个大大的赞！

[@shauniktaneja4733]

Thank you so much!

[@usmaven]

good explanation! better to put non-technical part first

[@amitkhandelwal2999]

Great Video. It would be great if you can also provide the info on how to deal with these concepts in practical scenario. I mean to say, how to increase power of test. How to decrease FP / FN / countereffects. That will give a complete end to end picture while dealing with them when someone encountered in such problems while implementing these things in practice. Loved all other videos which I have seen till today in your channel.

[@jaden2582]

could you explain the "AT LEAST as extreme as the data is actually observed" in the definition of the p value?

[@emma_ding]

Hey so an example would be when you are doing a test - if the means of two populations are the same, your null hypothesis is that those two are the same. Now you have observed data that shows that the difference is 1. “AT LEAST as extreme as the data is actually observed" means the difference is 1 or larger. 1 is the observed data and AT LEAST as extreme means that is the minimum difference. I hope this helps!

[@jaden2582]

@@emma_ding Thank you for this clear explanation!

[@qingchuanlyu4605]

This is really helpful. Now I know where my mistakes were!

[@mahdimerced]

Why did you delete most of the previous movies?

[@emma_ding]

You can find all my videos under the VIDEOS tab on my channel page. I changed the thumbnails of some videos a few weeks ago. :)

[@michellewww8036]

Like it!!!!!

[@yogiHalim]

95% confidence interval shows 95% from the center of a normal distribution population is represented. ie: 5% outliers are not represented by the equation

[@LouisChiaki]

A comment on the confidence interval, I think your interpretation (and a lot of data analyst) is from Frequentist's point of views. For Bayesian, there is no fixed true value.

[@nikhilmuthukrishnan7222]

You think your thumbnails are so cute!!! Well they are

[@InoHimeYa]

13 mins saves me at least 3 hours

[@robertwilsoniii2048]

This is really basic... how do jobs require multiple years of experience when these interview questions are just basic thing you learn in an intro stats class... ???

[@mussdroid]

#datascience

[@djjiang3718]

What a beautiful lady with high-quality content!

[@brothermalcolm]

Non-technical audience!

[@davidwarner1248]

Such a poor pronunciation

[@poopah4497]

the higher CL -> wider c.I? Is that a typo? I thought the opposite

[@emma_ding]

Hey Ruiruo! It's not a typo, the higher CL, the wider the CI, because increasing the confidence will increase the margin of error resulting in a wider interval.