Thanks for sharing Ranit! 👍 it’s good to also look at those courses on Udemy and EdX as well indeed. Can you share some of the courses you would recommend there?
@@Thuvu5 Yeah sure..... In Udemy, I particularly follow Jose Portilla's courses, he gives good overviews over Data Science topics plus he gives links to extra sources for theoretical portions. I've done my bachelor's in Statistics, so I find the series quite useful for me. In edX, Harvard provides course on Data Science course, and I found the syllabus much better than what IBM offers. MIT too provides course named Data Analytics Edge, with hands on exercises and tests.
@@ranitchatterjee5552 Woah this is so amazing, Ranit! Thanks so much for taking the time to suggest those courses 🙌🏽. I believe this will be very helpful for a lot of people looking to develop in those areas. I've pinned this comment so that more people can see it.
I am currently pursuing an edX certificate through Harvard called Data Analysis for Life Sciences (graduated with a biology degree looking into bioinformatics but I think I would like to go into general data scientist/analyst roles). I think it did a great job explaining the basic statistics and exploratory data analysis in R (better than the biostatistics class I took in college) but this certification only teaches how to use R so you probably want to get your knowledge on the ETL/ML knowledge somewhere else (I am starting the section of the certificate where they go over machine learning but it seems very brief and I hear Python is better for ML).
The problem is that everyone wants a shortcut to a formal education. In the alternative, they want to believe that hard work and grit will overcome gaps in knowledge. The solution: Do not rush into the job. Take your time and be methodical. Make sure you don't say to yourself "I don't really need to know that". Show respect to the industry. Be credible and do it right.
Totally agree, I'm currently taking a 4 month data science course and most of my classmates think that It's more important to learn pandas instead of statistics. But most of them didn't even have a basic programming or computer science knowledge before that; they just want to learn pandas, matplotlib, AWS and Docker, and to get a data science entry job after these 4 months
Hey there, thank you for watching! 👋Have you taken any of the Data analytics courses or certificates? What did you think about those courses? 🙂 I made a silly typo with "pratices" instead of "practices" at 0:32, I'm sorry for that 😅
Don't worry that's just a typo. You dont need to apologise. And yes I am just at the end of the Google Data Analytics Professional Certification and I agree they skipped the statistical analyses. Would love to have your ebook as you understand the real and minute struggles we face.
@@chhandomoydutta6397 Aw thank you!! And congrats on your progress with the cert.!! I'll definitely keep you posted on my the progress with the ebook. It's not always easy to find enough time to write but it will be coming! :)
Hi Thu Vu, the typo is no big deal. I am almost done with the Google Data Analyst Professional Certification Course and yes, there is little on statistics. Good video
I believe that those courses online courses are great, especially for beginners, but you need to read books and practice (make proyects as you mentioned in another video) to really master the software you're learning and filling gaps. I personally love Data Camp courses. I'm learning SQL, and the good thing about them is that you learn by coding, but the bad thing is that you don't even need to install the program you're learning. I love you're videos, keep it up!!!
Hey Eduardo, thanks for sharing this! I completely agree with you that online courses can only help that far. And indeed, the convenience of not having to install stuff when doing online course can be a pitfall as well. Thank you for watching and hope to see you in the next one! 🙌
One of the most commonly used techniques at work is the window /partition /row_number technique in SQL or Spark. Ive never seen it in any online courses
I was so confused for the last 2 months! Like there was a complete stop in my data journey but after watching this video I feel I have gathered some courage to go back and hit hard again! Thank you so much😊
Thank you for your advices. As a veteran who are just baby step to Data related field, i found your video is very informative as i am going through IBM course myself for starter. You put some ease in my mind knowing that there is a great youtube data advisor along my journey. thank you.
A difficulty I’ve found with IT training, whether online or face to face, in general is that often what is taught and the examples used depend more on what the writer finds easy to teach or thinks is cool, not what you might actually use in real life. I presume that the same holds true for data science courses.
As a current grad student in an Analytics program, I often find myself facing and wondering about the very same issues. I now have a clear understanding of where the gaps are and how I need to supplement my courses. You videos in general are resonating very strongly with many experiences I am facing as a beginner data scientist, I hope you continue the amazing work!
I’m entering my final months in receiving bachelor degree in data analytics and I have learned along the way that yes, there is no conventional standard way to teach how to code. However, a bad professor can simply make your intro to coding experience extremely depressing. There are a lot of nuances that have to do with understanding the very basic of programming, especially when you have never programmed. Example understanding what a notebook script like jupyter is to write code. Then understand how to even begin to use it, before you even begin to learn basics syntax of coding language. Coding is not difficult, what’s difficult is understanding the concepts behind to what it is you’re trying to achieve, and a good professor can at least structure the basics for you to set you up. In a nutshell, learn the basics of coding, using specific libraries for viz and modeling. Understand how to properly import your data and do EDA. once you get those basics, you then just have to practice on creating your own projects and analysis from them. Programming is never ending and there are tons of resources out there to help you with your code, you just have to learn to understand what the code tells you and input your own code into it to get your output results. of course you have to understand and know basic functions too. Anyway I am getting a headache just typing :). But good luck to everyone
Just found your channel, am I super happy I did. I'm currently taking the Google Data Analytics course and will be definitely learning these valuable skills in addition. Thank you
Thank you for the advice, I will be finished IBM DS Professional Certificate by the end of this month, after that, I think I will do some small DS projects on kaggle as soon as possible
thank you so much for pointing this out and recommending ways to practice what’s missing in courses, it’s so true, and actually all that you mentioned is what i look for in a course but they never get to teach those important topics! love your videos
I just came across your channel and I absolutely loved your video! Very funny but also informative!! I am a Senior Data Analyst and this was a great review for me!! Thanks
That's so nice of you to say, Connie! 👋 Glad it was informative and somewhat entertaining. I thought I might have tried too hard to make it funny 😂, but hey 🤷♀️
Dear Thu Vu, Thank you for making such comprehensively informative video -- you really have worked hard, especially while providing details in the section below. You really deserve likes and subscribes .. I am subscribing right away 😊
Hello Thu Vu, thank you for your sharing, what you said I totally agree about data science industry. I am still struggling everyday with learning how to code correctly, I hope one day we can share more about this industry.
Thank you for mentioning statistics! I’ve found statistical concepts are generally under-taught, which is a shame. For better or worse, “statistical significance” and “p-values” are generally understood by business stakeholders, which makes communication easier. Also basic stat techniques can probably solve 80% of all data science problems
How much value would you place on a Master’s Degree in Statistics? I have a government job, and I need to get a graduate degree to advance. Data Science graduate programs, even those offered by well known and established universities do not get me excited enough to drop tens of thousands of dollars.
Thank you I'm taking the google Data Analytics course. I noticed it talked a lot about past performance and results for present day but never really makes mention of how we should go about trying to forecast results through statistics.. Good to know I'm not crazy.
My learning method is courses, supplemented with as many projects as I can get my hands on. I download other peoples work, see how they’ve done things, attempt to break them, attempt to refine them and then apply their methods on other pieces of data and see how it performs and then make adjustments based on performance. For me, courses aren’t enough.. you have to jump in at the deep end, get lost and find a way out!
I am in the progress to change my career as data analyst, and your video helps me to point out what i should learn deeper. Thank you, it's awesome and i love it 👍
I have seen curriculums for several data science courses / academies, and I have the feeling that they try to sell the idea that you don't need any math for Data Science. Maybe because a lot of people think math is hard, and the courses sell better without a ton of math prerequisites? But in my humble opinion, mathematics is fundamental to data science, and more math is required than it's even taught at universities, let alone courses / academies. If you don't want to take my word for it, search for Foundations of Data Science by A. Blum, J. Hopcroft and R. Kannan. The book is available for free, just check the table of contents :D
Thanks a lot for sharing the book! I just looked at it, great stuff!! Agreed with you that many DS courses overlook the math part. I often have to look elsewhere to learn the necessary math. I’m going to talk about this in one of the coming videos! 🙌
@@Thuvu5 Another great book is Mathematics for Machine Learning, it's also available for free online. It's a great place to start if you want to see what math is needed for machine learning, it also discusses how that math is applied in machine learning. Maybe you can mention / link the books in your video so other people can get them too.
Thanks a lot Thu Vu for your insightful expose of the weaknesses that some of these courses have. Books have always been looked down upon but I realize they are quite helpful more than online courses especially when you get the right one. I have been studying Inferential Statistics and Introduction to Statistics in a book called Introduction to Statistical Thinking (With R, Without Calculus) by Benjamin Yakir, The Hebrew University. The book provides more profound info on R in a way you won't find in most online courses.
Oh thank you so much for the book recommendation!! It’s a great book, I’ll recommend it to others as well. Totally agreed with you and the value of using books in complement with the courses 🙌
Just discovered your channel!! Three vids in a row, i love the way you expose topics and tips about how to get good practice. I'm an actuary but DS is more interesting!!
Aw thanks so much Luis! So glad you find the vids useful and binge-watch them 😄. I’m also working in actuary at work but I only focus on DS stuff, because it’s indeed more fun 😏🙌
@@Thuvu5 Definitely you're right. I'd love to make the transition to the DS tasks in a near future. It's a big help the resources from your channel. May I ask you, what statistics textbook do you consider it's a good support for the DS or what you reccommend me, from theoretical viewpoint or python oriented? Thanks again!! And keep it up that way!
@@luisdmoreyra Hi Luis, the book Introduction to Statistical learning in R (link in the video description) is definitely a good book to start with. It's an amazing book, it's in R tho. For Python I haven't found an equivalent book to this unfortunately.. But this one Hands-on Machine Learning with Scikit-Learn, Keras, and TensorFlow is the one I have used and would recommend for practical coding and machine learning. Hope this helps! :)
Thank you for your great video. But It is worth noting that multicollinearity is not an assumption of linear regression, it can still impact the reliability and interpretation of the model's coefficients. However, despite the presence of multicollinearity, a linear regression model can still provide reasonably accurate predictions for the output variable, which in this case is salary. I also agree that It is also essential to carefully consider the interpretation of the coefficients and their associated standard errors in order to draw meaningful conclusions from the model.
Denying the applicability of linear regression when multicollinearity exists is a "blind" :) approach. It disregards the vast range of scenarios where linear regression can be utilized, essentially weakening one's most powerful tool.
Exactly the same questions were storming through my mind as I am quite new to DS, I was loosing my confidence for the interview, but now i know that's the norm. A sigh of relief. Peace
I don't know how you showed up as a TH-cam recommendation. I have been learning some web programming lately. I'm new to the whole programming thing. It would be nice if you had a video on why a web programmer might be interested in learning data science. I'm all over the place with this coding thing so some guidance would be really appreciated. Your content looks good so I will subscribe
Aww I’m glad you found my channel 👋. Your suggestion sounds a nice idea, I’ll consider that for future videos. Thanks a lot for suggesting! All the best for you with learning 💪
I agree with most of the observations in this video. Where I disagree is modularisation of functions. ingestiong +cleaning + transformation could be put in the same wrangling function. This same wrangling function can be used for different data sources that require the same process. Otherwise you find yourself calling multple functions in similar sequence to do same thing on different sources of data i.e. duplicating code and making it harder to read for third parties. I think docstrings and function annotations and discreptive naming should help navigating the code.
Thank you very much, this has been one of the most usefull videos i've seen on the topic of DS. You've definitely earned yourself a new subscriber and I will watch more of your videos on the topic.
I'm trying to improve my English by watching videos like yours or taking courses in DataCamp, this year I decided to change of specialty from Frontend developer to Data Scientist, I recognize that the Statistic was difficult in the beginning. I finished the BootCamp in LeWagon recently and now I continue studying while I looking for my first job in data science. Good video!
That's amazing Alex! I think Frontend developer is cool too tho 🤓. Good luck with learning and the job search! 🍀 And don't forget to enjoy the journey :)
Honestly i clicked on the video because of tumbnail, but I don't regret it. The issues you addressed in this video are essential. Most of the online courses don't teach this. And from my personal experience git is must. I will definately checkout more videos from your channel.
Hi Saurabh, totally agree with you that knowing Git is a must! When I started learning about Git at my work I thought, how the hell I didn't learn about this before.. Thanks for your support and hope you'll enjoy other videos too! 😀
I agree that some courses doesn't cover all the things. Everyone neglects the fact that data science is easy for those who are good with statistics and algebra. Sometimes you need to learn things by yourself. 😁
Agree with the take on using books because they provide nuances that videos don't moreso for statistics. After getting a book , everything that I watched on videos started to make alot more sense
For linear and multivariate regression introductory and intermediate econometrics 101 tyoe books with code is highly recommended. Specially the framework or viewpoint of how you use statistics. Finding and processing data sets never easy and very difficult often. But it just starts there, knowing what they imply to your specific use case not just standard statistical description could be critical. The difference between observation and understanding / insights that in business leads to strategic decisions.
Từ trước đến giờ khi em tự học python thì chưa thấy ai chỉ chi tiết để viết code cho dễ đọc và hợp lí cả. Cảm ơn những chia sẻ rất hữu ích của chị. Mà video còn rất vui nữa.
this video describes exactly what i'm passing now at my first job as a data scientist !!!! :) there are too many 'gits' !!! gitlab, git, smartgit... why ?? I still feel butterflies in my stomach before doing a commit/push thinking it will cause some problem lol ...... I feel a little bad and sometimes worried that I'm learning everything too quickly or in a 'run over' way... it's always that feeling of building the plane while it's flying
Haha great to hear you could relate to that 😁. I agree that it feels like building the plane while it's flying. So many things to learn and especially we have to deliver things at work while still figuring things out on the fly. I definitely passed many aha and "ugh? wtf" moments in my data science career 😂. Yeah, the git thing indeed confused the heck out of me as well haha. Gitlab, GitHub, bitbucket, etc. are just different hosting services for Git. The worries of doing something wrong with Git definitely adds some thrill at work 🤣🙌🏽
Jetbrains teaches you about good code practices in their courses. They have Java, Kotlin and Python. What I've learned from many courses doesn't come close to the Jetbrains approach, it focuses on best practices, code readability structure, and teaching how to think outside the box.
Hey Guillaume, yes is still an ongoing idea, but I’m struggling to make time for writing. I’m also considering making a live course instead. Let me know what you think!
@@Thuvu5 that sounds pretty good, perhaps the ebook can be a transversal work throughout the writing of the courses ? In any case keep it up, the format is really enjoyable, not too much information, but still really informative !
They also don't teach you methods to gather the data you want to analyse, how to design your research study, how to work with live data stream on the fly, ways to deploy, automate and maintain your reports/visualizations, where potentially you can apply data science for. They're more like a guide to a test question like: "Given the data set bellow, find the appropriate function to correlate the variables and plot the output on a chart"
0:30 onwards, yeah, code style is not taught in so many courses, i am glad i took the more thorough and lengthy and rigorous fundamental materials, but on the plus side, they covered these very well 3:50 thanks a lot for mentioning the actual courses too, it helps a lot 4:455:31 that example really helped in understanding the importance of that 8:26 _code vectorisation and lambda functions make code faster_ lambda functions = functional programming, hmm, so glad that i just watched the computerphile vid on it yesterday 8:34 _... and measuring it by timeit_ - nice 12:05 _on tableau_ - ohkay, so, tableau is used for those dashboards, got it
@@Thuvu5 8:34 glad i had shared this timestamp here before. i was searching for exactly this today. basically, i was watching other video, and they used time.time() as a wrapper and a decorator, then someone said to use time.perf_counter() but now, the %timeit cell magic or say the timeit.timeit() function seems to be the best - with repetition number and everything thanks again 😃 manuals link: * docs python library timeit * ipython readthedocs interactive magics * docs python library time
I agree with every word you have said and I've encountered same experience with these Coursera courses. Udemy and edx courses are more reliable.
Thanks for sharing Ranit! 👍 it’s good to also look at those courses on Udemy and EdX as well indeed. Can you share some of the courses you would recommend there?
@@Thuvu5 Yeah sure..... In Udemy, I particularly follow Jose Portilla's courses, he gives good overviews over Data Science topics plus he gives links to extra sources for theoretical portions. I've done my bachelor's in Statistics, so I find the series quite useful for me. In edX, Harvard provides course on Data Science course, and I found the syllabus much better than what IBM offers. MIT too provides course named Data Analytics Edge, with hands on exercises and tests.
@@ranitchatterjee5552 Woah this is so amazing, Ranit! Thanks so much for taking the time to suggest those courses 🙌🏽. I believe this will be very helpful for a lot of people looking to develop in those areas. I've pinned this comment so that more people can see it.
Could you guys provide links ? Sorry being lazy here
I am currently pursuing an edX certificate through Harvard called Data Analysis for Life Sciences (graduated with a biology degree looking into bioinformatics but I think I would like to go into general data scientist/analyst roles). I think it did a great job explaining the basic statistics and exploratory data analysis in R (better than the biostatistics class I took in college) but this certification only teaches how to use R so you probably want to get your knowledge on the ETL/ML knowledge somewhere else (I am starting the section of the certificate where they go over machine learning but it seems very brief and I hear Python is better for ML).
5 Things DS Courses Don't Teach You
0:30 - (1) Coding Readability
2:44 - (2) Data Visualization Skills / Communication / Storytelling
3:42 - (3) Statistical Skills
6:55 - (4) Collaboration Tools (i.e. Git Version Control)
7:41 - (5) Coding Performance
The problem is that everyone wants a shortcut to a formal education. In the alternative, they want to believe that hard work and grit will overcome gaps in knowledge. The solution: Do not rush into the job. Take your time and be methodical. Make sure you don't say to yourself "I don't really need to know that". Show respect to the industry. Be credible and do it right.
Well said 👏
Totally agree, I'm currently taking a 4 month data science course and most of my classmates think that It's more important to learn pandas instead of statistics. But most of them didn't even have a basic programming or computer science knowledge before that; they just want to learn pandas, matplotlib, AWS and Docker, and to get a data science entry job after these 4 months
@@rknapx Their lack of dedication will become evident at some point during their career.
Hey there, thank you for watching! 👋Have you taken any of the Data analytics courses or certificates? What did you think about those courses? 🙂
I made a silly typo with "pratices" instead of "practices" at 0:32, I'm sorry for that 😅
Don't worry that's just a typo. You dont need to apologise. And yes I am just at the end of the Google Data Analytics Professional Certification and I agree they skipped the statistical analyses. Would love to have your ebook as you understand the real and minute struggles we face.
@@chhandomoydutta6397 Aw thank you!! And congrats on your progress with the cert.!! I'll definitely keep you posted on my the progress with the ebook. It's not always easy to find enough time to write but it will be coming! :)
Hi Thu Vu, the typo is no big deal. I am almost done with the Google Data Analyst Professional Certification Course and yes, there is little on statistics.
Good video
Loved the skits throughout this! 🙌🏼😂
Also thanks for the shoutout!!
Aww, I'm so glad you found it funny 🤣. I was a bit nervous about it haha. You're welcome!! Thanks so much for watching 🙌🏽
Thanks!
Was researching this topic and this video was super helpful, thank you!
I'm so glad it helped with your research, Shane! Thanks for watching! 👋
One of the more helpful data analytics overviews out there. Great resources attached as well. Thank you!
I believe that those courses online courses are great, especially for beginners, but you need to read books and practice (make proyects as you mentioned in another video) to really master the software you're learning and filling gaps. I personally love Data Camp courses. I'm learning SQL, and the good thing about them is that you learn by coding, but the bad thing is that you don't even need to install the program you're learning.
I love you're videos, keep it up!!!
Hey Eduardo, thanks for sharing this! I completely agree with you that online courses can only help that far. And indeed, the convenience of not having to install stuff when doing online course can be a pitfall as well. Thank you for watching and hope to see you in the next one! 🙌
One of the most commonly used techniques at work is the window /partition /row_number technique in SQL or Spark. Ive never seen it in any online courses
That's true! It's so strange.. Thanks for sharing this observation, Sam!
I was so confused for the last 2 months! Like there was a complete stop in my data journey but after watching this video I feel I have gathered some courage to go back and hit hard again! Thank you so much😊
Best realistic advice I've heard in six years. Thank you.
Hi Thu Vu, please do make the e-book it would be really educational for beginner data analysts and you are spot on with info 👌🏽
You're definitely my favorite TH-camr.
I'm on my path to become a Data Analyst Fresher. Your videos have greatly influenced me. Thank you very much!
So glad to hear this 🤗. Thank you for watching, Tung!
I’m at the beginning of my Data Science cert and I’m excited I found this vid! Looking forward to reviewing your recommended resources soon
Glad it was useful! Thanks for watching Anita! Good luck with your study 🍀
Golden insights, thanks for sharing that!
Thank you for your advices. As a veteran who are just baby step to Data related field, i found your video is very informative as i am going through IBM course myself for starter. You put some ease in my mind knowing that there is a great youtube data advisor along my journey. thank you.
Aw thank you for such kind words! I’m really glad my content is useful for you. Good luck with your learning 🍀
A difficulty I’ve found with IT training, whether online or face to face, in general is that often what is taught and the examples used depend more on what the writer finds easy to teach or thinks is cool, not what you might actually use in real life. I presume that the same holds true for data science courses.
As a current grad student in an Analytics program, I often find myself facing and wondering about the very same issues. I now have a clear understanding of where the gaps are and how I need to supplement my courses. You videos in general are resonating very strongly with many experiences I am facing as a beginner data scientist, I hope you continue the amazing work!
I would be really interested in that ebook! Your videos are always so helpful!
I’m entering my final months in receiving bachelor degree in data analytics and I have learned along the way that yes, there is no conventional standard way to teach how to code. However, a bad professor can simply make your intro to coding experience extremely depressing. There are a lot of nuances that have to do with understanding the very basic of programming, especially when you have never programmed. Example understanding what a notebook script like jupyter is to write code. Then understand how to even begin to use it, before you even begin to learn basics syntax of coding language. Coding is not difficult, what’s difficult is understanding the concepts behind to what it is you’re trying to achieve, and a good professor can at least structure the basics for you to set you up. In a nutshell, learn the basics of coding, using specific libraries for viz and modeling. Understand how to properly import your data and do EDA. once you get those basics, you then just have to practice on creating your own projects and analysis from them.
Programming is never ending and there are tons of resources out there to help you with your code, you just have to learn to understand what the code tells you and input your own code into it to get your output results. of course you have to understand and know basic functions too.
Anyway I am getting a headache just typing :). But good luck to everyone
Just found your channel, am I super happy I did. I'm currently taking the Google Data Analytics course and will be definitely learning these valuable skills in addition. Thank you
Very good overview, thanks!
Thank you for the advice, I will be finished IBM DS Professional Certificate by the end of this month, after that, I think I will do some small DS projects on kaggle as soon as possible
Hey, that sounds awesome! 👍 You can check out some other data projects on my channel, hope you might find something useful 😊
Thank a lot for your insightful sharing. These tips definitely improve my work
Very well said Thu. 100% agree
Awesome video..i encountered similar issues while learning data science courses
thank you so much for pointing this out and recommending ways to practice what’s missing in courses, it’s so true, and actually all that you mentioned is what i look for in a course but they never get to teach those important topics! love your videos
Thank you so much for your kind comment, Rox! So glad it was helpful 🙌🏽
It would love to read your ebook. Experience is so valuable in this field. Thank for your great information
thanks for the great video and looking forward to the ebook
Hey, thank you for commenting! Okay I’d better get started with writing now 😂
You have very good points on your video, well done.
Thanks Sebastian! So glad you enjoyed it 👋
Wow great comprehensive and funny video. Thank you Thu!
Glad you enjoyed it, Gallow!! 🙌
I love this. Thank you so much. I'm currently learning data analysis and this helps A LOT
Hey, that's awesome - good luck with your learning 💪. Glad it helped!
That's all what I also felt. Great contents!
Yeah.. glad to hear I was not the only one to feel this. Thanks for tuning in, Kaira! 👋
You're amazing! Thank you!
Girl tht's amazing thanks for the tips
I just came across your channel and I absolutely loved your video! Very funny but also informative!! I am a Senior Data Analyst and this was a great review for me!! Thanks
That's so nice of you to say, Connie! 👋 Glad it was informative and somewhat entertaining. I thought I might have tried too hard to make it funny 😂, but hey 🤷♀️
Dear Thu Vu,
Thank you for making such comprehensively informative video -- you really have worked hard, especially while providing details in the section below. You really deserve likes and subscribes .. I am subscribing right away 😊
Aw that’s so kind of you, Aamir!! Thanks for watching and subscribing 🙌😃
I’ve been thinking about this a lot lately! This video came at a great time!
Hey Emmanuel, thank you for watching!!👋 Glad to hear it’s relevant for you! 😊
Hello Thu Vu, thank you for your sharing, what you said I totally agree about data science industry. I am still struggling everyday with learning how to code correctly, I hope one day we can share more about this industry.
Thank you for mentioning statistics! I’ve found statistical concepts are generally under-taught, which is a shame. For better or worse, “statistical significance” and “p-values” are generally understood by business stakeholders, which makes communication easier. Also basic stat techniques can probably solve 80% of all data science problems
True
How much value would you place on a Master’s Degree in Statistics? I have a government job, and I need to get a graduate degree to advance. Data Science graduate programs, even those offered by well known and established universities do not get me excited enough to drop tens of thousands of dollars.
LOL great acting skills!!! Way to go Thu Vu!!!! ROFL!
Very good. +1 for learning statistical methods.
I love your videos! They have a lot of cozy, funny and useful moments. Thanks a lot for all this information that you share with us!
Thank you so much Sophia!! Really appreciate your kind words 💜
Thanks for the tips. This has been really helpful.
I'd buy your ebook! Your videos are so unique and informative. Keep it up!
Hey Andrey, that’s so kind of you to say, I really appreciate that! 😄 Thanks for watching and commenting! 🙌
Great video, I agree with everything.
Amazing video!
I'm just starting my way to data science and this video was super insightful!!!
Yaaay I’m so glad you found it useful! Have fun learning!! 😄🙌
Amazing! Thank youu!!
Thank you for sharing.
Thank you for making these videos! They are really helpful! 😄
Thank you I'm taking the google Data Analytics course. I noticed it talked a lot about past performance and results for present day but never really makes mention of how we should go about trying to forecast results through statistics.. Good to know I'm not crazy.
This was a great video. IO also agree with every word you've said.
My learning method is courses, supplemented with as many projects as I can get my hands on.
I download other peoples work, see how they’ve done things, attempt to break them, attempt to refine them and then apply their methods on other pieces of data and see how it performs and then make adjustments based on performance.
For me, courses aren’t enough.. you have to jump in at the deep end, get lost and find a way out!
Totally agree with you R! Thanks for sharing your experience! 🙌
Subscribed, Great job. Absolutely right about everything.
Yaaay! Thanks for watching and subscribing Oluwadara! 🙌
Looking forward to your E-book!
I am in the progress to change my career as data analyst, and your video helps me to point out what i should learn deeper. Thank you, it's awesome and i love it 👍
Hey, thanks a lot for commenting and all the best with your progress! 👋
It's gonna be a good idea to write a e-book as a guide for as, you already help me so much 😊 ❤️ thank you
Great video, simple and to the point. Keep sharing🙌
That was very informative
Great video I have ever seen so far.
I have seen curriculums for several data science courses / academies, and I have the feeling that they try to sell the idea that you don't need any math for Data Science. Maybe because a lot of people think math is hard, and the courses sell better without a ton of math prerequisites? But in my humble opinion, mathematics is fundamental to data science, and more math is required than it's even taught at universities, let alone courses / academies. If you don't want to take my word for it, search for Foundations of Data Science by A. Blum, J. Hopcroft and R. Kannan. The book is available for free, just check the table of contents :D
Thanks a lot for sharing the book! I just looked at it, great stuff!! Agreed with you that many DS courses overlook the math part. I often have to look elsewhere to learn the necessary math. I’m going to talk about this in one of the coming videos! 🙌
@@Thuvu5 Another great book is Mathematics for Machine Learning, it's also available for free online. It's a great place to start if you want to see what math is needed for machine learning, it also discusses how that math is applied in machine learning. Maybe you can mention / link the books in your video so other people can get them too.
Thanks a lot Thu Vu for your insightful expose of the weaknesses that some of these courses have. Books have always been looked down upon but I realize they are quite helpful more than online courses especially when you get the right one. I have been studying Inferential Statistics and Introduction to Statistics in a book called Introduction to Statistical Thinking
(With R, Without Calculus) by Benjamin Yakir, The Hebrew University. The book provides more profound info on R in a way you won't find in most online courses.
Oh thank you so much for the book recommendation!! It’s a great book, I’ll recommend it to others as well. Totally agreed with you and the value of using books in complement with the courses 🙌
Thank you for the resources
Just discovered your channel!! Three vids in a row, i love the way you expose topics and tips about how to get good practice. I'm an actuary but DS is more interesting!!
Aw thanks so much Luis! So glad you find the vids useful and binge-watch them 😄. I’m also working in actuary at work but I only focus on DS stuff, because it’s indeed more fun 😏🙌
@@Thuvu5 Definitely you're right. I'd love to make the transition to the DS tasks in a near future. It's a big help the resources from your channel. May I ask you, what statistics textbook do you consider it's a good support for the DS or what you reccommend me, from theoretical viewpoint or python oriented? Thanks again!! And keep it up that way!
@@luisdmoreyra Hi Luis, the book Introduction to Statistical learning in R (link in the video description) is definitely a good book to start with. It's an amazing book, it's in R tho. For Python I haven't found an equivalent book to this unfortunately.. But this one Hands-on Machine Learning with Scikit-Learn, Keras, and TensorFlow is the one I have used and would recommend for practical coding and machine learning. Hope this helps! :)
Thank you for your great video. But It is worth noting that multicollinearity is not an assumption of linear regression, it can still impact the reliability and interpretation of the model's coefficients. However, despite the presence of multicollinearity, a linear regression model can still provide reasonably accurate predictions for the output variable, which in this case is salary. I also agree that It is also essential to carefully consider the interpretation of the coefficients and their associated standard errors in order to draw meaningful conclusions from the model.
Sorry correction: I mean No multicollinearity is not an assumption for linear regression
Denying the applicability of linear regression when multicollinearity exists is a "blind" :) approach. It disregards the vast range of scenarios where linear regression can be utilized, essentially weakening one's most powerful tool.
ciaooo
i have just open "an introduction to statistical learnig"
thank you 4 this advice!
Thanks for the video. Very useful.
You are really creating some quality content! Thanks again and you are both smart and pretty.
Thank you so much! ☺️
Exactly the same questions were storming through my mind as I am quite new to DS, I was loosing my confidence for the interview, but now i know that's the norm. A sigh of relief. Peace
Thank you so much!
I don't know how you showed up as a TH-cam recommendation. I have been learning some web programming lately. I'm new to the whole programming thing. It would be nice if you had a video on why a web programmer might be interested in learning data science. I'm all over the place with this coding thing so some guidance would be really appreciated. Your content looks good so I will subscribe
Aww I’m glad you found my channel 👋. Your suggestion sounds a nice idea, I’ll consider that for future videos. Thanks a lot for suggesting! All the best for you with learning 💪
I love u, the only way to learn is being curious and hacking everywhere :)
Thank you very much!
I agree with most of the observations in this video. Where I disagree is modularisation of functions. ingestiong +cleaning + transformation could be put in the same wrangling function. This same wrangling function can be used for different data sources that require the same process. Otherwise you find yourself calling multple functions in similar sequence to do same thing on different sources of data i.e. duplicating code and making it harder to read for third parties. I think docstrings and function annotations and discreptive naming should help navigating the code.
Thank you very much, this has been one of the most usefull videos i've seen on the topic of DS. You've definitely earned yourself a new subscriber and I will watch more of your videos on the topic.
Yaaay! That’s so awesome 🤩. Thanks for tuning in and subscribing! 🙌
I can relate to the Git part. I emailed an entire folder.
Great video!
I'm trying to improve my English by watching videos like yours or taking courses in DataCamp, this year I decided to change of specialty from Frontend developer to Data Scientist, I recognize that the Statistic was difficult in the beginning. I finished the BootCamp in LeWagon recently and now I continue studying while I looking for my first job in data science.
Good video!
That's amazing Alex! I think Frontend developer is cool too tho 🤓. Good luck with learning and the job search! 🍀 And don't forget to enjoy the journey :)
I think you are my big heroine from my life.
I hope to know if you finished making your ebook!
I love your all of informaiton. Thank you again❤️
Honestly i clicked on the video because of tumbnail, but I don't regret it. The issues you addressed in this video are essential. Most of the online courses don't teach this. And from my personal experience git is must. I will definately checkout more videos from your channel.
Hi Saurabh, totally agree with you that knowing Git is a must! When I started learning about Git at my work I thought, how the hell I didn't learn about this before.. Thanks for your support and hope you'll enjoy other videos too! 😀
This is gold.!
Thank you 🙌
Angel speaks about data science :)
I agree that some courses doesn't cover all the things.
Everyone neglects the fact that data science is easy for those who are good with statistics and algebra.
Sometimes you need to learn things by yourself. 😁
Agree with the take on using books because they provide nuances that videos don't moreso for statistics. After getting a book , everything that I watched on videos started to make alot more sense
Totally agree, Cheryl!! 🙌 I’ve had exactly the same experience
Absolutely love your channel!
Hi Fleur, thank you so much! I'm so glad to hear that 🥰
great content as always
i've fallen in love with your channel
Aww thank you for this!! 🥰
For linear and multivariate regression introductory and intermediate econometrics 101 tyoe books with code is highly recommended. Specially the framework or viewpoint of how you use statistics. Finding and processing data sets never easy and very difficult often. But it just starts there, knowing what they imply to your specific use case not just standard statistical description could be critical. The difference between observation and understanding / insights that in business leads to strategic decisions.
Từ trước đến giờ khi em tự học python thì chưa thấy ai chỉ chi tiết để viết code cho dễ đọc và hợp lí cả.
Cảm ơn những chia sẻ rất hữu ích của chị. Mà video còn rất vui nữa.
very great video!
this video describes exactly what i'm passing now at my first job as a data scientist !!!! :) there are too many 'gits' !!! gitlab, git, smartgit... why ?? I still feel butterflies in my stomach before doing a commit/push thinking it will cause some problem lol ...... I feel a little bad and sometimes worried that I'm learning everything too quickly or in a 'run over' way... it's always that feeling of building the plane while it's flying
Haha great to hear you could relate to that 😁. I agree that it feels like building the plane while it's flying. So many things to learn and especially we have to deliver things at work while still figuring things out on the fly. I definitely passed many aha and "ugh? wtf" moments in my data science career 😂. Yeah, the git thing indeed confused the heck out of me as well haha. Gitlab, GitHub, bitbucket, etc. are just different hosting services for Git. The worries of doing something wrong with Git definitely adds some thrill at work 🤣🙌🏽
Jetbrains teaches you about good code practices in their courses.
They have Java, Kotlin and Python.
What I've learned from many courses doesn't come close to the Jetbrains approach, it focuses on best practices, code readability structure, and teaching how to think outside the box.
Which one would you recommend:
1. Datacamp
2. Dataquest
3. LinkedIn Learning Path
4. Coursera
5. Kaggle
6. TH-cam
I would personally love to read your E-Book
I'd love to read your hitchhiker e-book, if that is still an ongoing project / idea !
Hey Guillaume, yes is still an ongoing idea, but I’m struggling to make time for writing. I’m also considering making a live course instead. Let me know what you think!
@@Thuvu5 that sounds pretty good, perhaps the ebook can be a transversal work throughout the writing of the courses ?
In any case keep it up, the format is really enjoyable, not too much information, but still really informative !
Thank you 🙏🙏
They also don't teach you methods to gather the data you want to analyse, how to design your research study, how to work with live data stream on the fly, ways to deploy, automate and maintain your reports/visualizations, where potentially you can apply data science for.
They're more like a guide to a test question like:
"Given the data set bellow, find the appropriate function to correlate the variables and plot the output on a chart"
Indeed, these are very interesting points, Marcos!
0:30 onwards, yeah, code style is not taught in so many courses, i am glad i took the more thorough and lengthy and rigorous fundamental materials, but on the plus side, they covered these very well
3:50 thanks a lot for mentioning the actual courses too, it helps a lot
4:45 5:31 that example really helped in understanding the importance of that
8:26 _code vectorisation and lambda functions make code faster_
lambda functions = functional programming, hmm, so glad that i just watched the computerphile vid on it yesterday
8:34 _... and measuring it by timeit_ - nice
12:05 _on tableau_ - ohkay, so, tableau is used for those dashboards, got it
Thank you for sharing these experiences Yash! 😀💯
@@Thuvu5
8:34 glad i had shared this timestamp here before. i was searching for exactly this today.
basically, i was watching other video, and they used time.time() as a wrapper and a decorator,
then someone said to use time.perf_counter()
but now, the %timeit cell magic or say the timeit.timeit() function seems to be the best - with repetition number and everything
thanks again 😃
manuals link:
* docs python library timeit
* ipython readthedocs interactive magics
* docs python library time