Great video, I'm a math/stat major however I feel like I'm not good enough🙃 I will be taking analysis in the near future and im really scared 😔 I feel like I don't belong in the math world. Anyhow, for anyone looking to improve just know that there is no easy road to understanding math. A serious amount of work and dedication are needed to succeed🤓 although the rewards are definitely worth it... Best of luck
Linear algebra by Gilbert Strang is the best material available to study. At the beginning, I didn't have patience to complete it but in the end, I realized this is one of the finest Math people in our current world.
The content on this channel is incredibly good. This has become one of my favorite channels. Your analytical attitude and your optimism is endearing. Keep up the great work!
5 days ago: Watched Joma's why you NEED math Today: Watched Tina's How to learn math Great video Tina! Awesome breakdown of how to approach learning math for data science!
Really love how you are so upfront and generous about sharing your background! A lot of others seem to say, "I was good at math", or "I loved math ever since", so it's kind of difficult for aspiring DS who don't have strong backgrounds in math, to relate. But I'm so glad you shared this! It is really inspiring amd motivating. Thank you 💕
The best way to learn datascience is keep doing personal projects, kaggle tournaments, and read lots interesting fun papers in ML/DL. Turn it into your hobby.
Hi Tina, this video is super helpful! similar to your experience, I also dropped out of my Linear Algebra class last spring and retook it in summer (even extending my graduation!). The reason why I dropped out is that I was so confused by the proofs and was literally failing the midterm. But when I retake it again, I was able to identify my weaknesses, and felt more comfortable in the material, ending up with a good grade. I also recommend everyone who panics about math instinctively to give it another chance or a little more time!
I am a Full Stack Developer with over 25+ years of development experience. I am currently teaching myself data science. With that being said, I have always wondered why math is important, after googling a bunch of videos, yours was the first one to answer the question that I was looking for. And yes, I have always sucked at math, but for some reason as I have gotten older, I like math. Thanks for this very informative video, I have learned Linear Regression, RandomForest and Artificial Nueral Network algorithms so a lot of what you were saying I was familiar with. But the point you made about not necessarily knowing how to solve algebra problems, but rather understanding the meaning behind the problem hit it on the nail on the head. Thanks for a great video.
You know I never thought about math that way?? I took Calculus and Linear Algebra in undergrad and graduate school. But now after hearing your explanations, I’m going to pull out those math books and rework some of those math problems, and keeping in mind what you have said. Thanks!
I really like this learning approach of a relentless focus on projects. What frustrates me the most about learning is how I can never retain everything I learn. At the end of the day you mostly only retain what you put into practice, so why not do all your learning at the same time as putting it into practice on projects? It just makes sense really.
this is so helpful. I am also super good with bio but once the class was about math I start to panic. As an engineer student it was painful and I felt so ashamed of myself among those genius kids. And this really hindered my decision when I chose lab, I just automatically shyed away from any thing related to C'S and calculation even though I was interested in them. Thank you Tina, now I am picking up programming again
thank you for being honest and open about your math background, this is very helpful. I am filling in the gaps of my knowledge backwards, I know how to code and make sklearn models but i don't have the understanding behind me.
I didn’t think you would talk about the fact that most ml algorithms are simply related to distances to compare how far away our prediction is from the ground truth, but you did at 5:20!!!! Very nice 👍
This is an incredible video! Thank you for sharing your experience, I can relate to the paralyzing fear/panic with math sooo hard hahaha Math is so much about confidence I feel. I used to love it (still do honestly) when I was younger & was actually really good at it. But somewhere along the way I lost my confidence & self-doubt started creeping in after one too many sucky exams lol BUT I still love math (even though I am not as good at it as I’d like) and am pursuing a stem major now - not giving up! 😄
You make really nice videos which inspire me while upscaling and searching for a job a year after completing my Master's in Data Analytics in a foreign country during Covid-19. You have a very calm and friendly way of educating others. Yet, for some reason though I find it weird that the eq is y=b1x+b0 instead of y=mx+c. But that's just me... Kudos!
Thank you so much for your kind words 😊 and haha I was debating putting that form but the way I kinda see it in ML is usually the same letter and subscripts because they are all coefficients.
I was good at Maths during my High School because that is the only thing I have ever loved and never get bored of solving that maths problem before programming.
Working as a data scientist for a while now, I can attest to every bit you said. Like you, I pursued my undergrad degree in one of the biological sciences (Pharmaceutics) but it turned out to be a 'no' for me and am glad I pivoted to this field. In many instances either its the complexity of the math or the wish to jump directly to the fancy stuff that keeps people from delving deeper into the inner workings of the algorithm. I cannot emphasize enough how vital it is to know at least something about how things work under the hood. Sure it takes time, persistence and effort to have that kind of knowledge but its always good for both short and long run. So just my two cents of unsolicited advice to budding analysts, data scientists, statisticians etc. - scare the math scare away and be willing to learn day in and day out and it will be for sure a much comfortable ride down the lane. Btw great videos! Keep going.
Math was my favorite subject in high school but I am TERRIBLE at it. I was intimidated in college by the people that just immediately understood a concept and the professor would move on so quickly. I changed majors due to this. My husband is in this field and It has made me want to join it. I think self paced learning and mentorship with my husband will help me with this.
good video! minor correction: normally least square is minimizing vertical distance, instead of perpendicular distance. i.e. the plot at 3:18 is perpendicular distance:)
linear regression don't do well when :- -relationship b/w independent and dependent is non linear. - affected by outliers - prone to overfitting - prone to multicollinearity - Inability to determine Feature importance
Tina, this is an excellent intuitive explanation of how the least squares method works, but it's not 100% accurate. It minimizes the vertical distance (for each x, y-hat(x) - y(x)) - the lines should be perpendicular to the x-axis, not the regression line. So in your diagram, just draw the lines up/down from each point until they hit the regression line and you have a perfect and accurate description of ordinary least squares.
I reckon math is as like other subjects, but the thing is that it's not written in english, so our brain says that hey this is a unique thing + as it's not in english well that's hard then. I think that's the only reason why people don't really understand maths, otherwise maths is also easy as other subjects are or any other thing to learn. I also used to think that it's hard but once I thought that I should really give it a try and let's se how it works for me(at least 3 hours a day, even if becomes a hectic) and after 3 or 4 days of continues learning it, it turns out that yes It's easy and I can learn it pretty easily like any other subject.
Im a psychologist from Germany. We have quite heavy stats training and do a lot of modeling. I understand everything our analysts talk about when I go into their meetings but can't imagine companies hiring psychologists as data scientists. Have you encountered psychologists that worked as data scientists?
Hmmm I personally have not, although I have seen people in user research role that are psychologists. I don’t see why you can’t be a data scientist though if you are willing to do some data science projects + some networking.
I found that linear regression is weaker than other algorithms at reducing error rates and does poorly if the data set has outliers. It can, however, predict outcomes based on at least one variable of a changing value. Stellar video.
I’m from Finance and Economics background and I am planning to step out my comfort zone to take MSc in Data Science and Business Analytics. Hopefully I can graduate smoothly lol
Thanks a lot Tina for the video. Can you make us a video of how can we work remotely Full - time from home as Data Scientists ? and how much money can we make ?
The reason why so many intelligent people struggle with math is that math education is broken. It focuses on teaching people how to use math instead of how to do math. They are taught the solutions to problems and then made to apply that already formed solution to the same problem with different details. The result is that math education is completely divorced from the field, and students do not view themselves as mathematicians. They were not given the mental tools of logic to solve problems. They were given the solutions to problems. The quadratic formula is not a method to solve a problem, it is the answer to a problem. We give them the answer to questions. We don't teach them how to form solutions to new problems. That's why so many kids, upon learning some formula for solving some absurdly specific problem, ask "when am I ever gonna need this?" And they're right.
I'm only half kidding here but an asian saying "Math has always been my weakest subject" has given me so much confidence lol. I'm learning math now from algebra 1 on Khan Academy and I plan to work my way through everything up to calculus after which I'll do stats and then start to learn data science then refer back to the math I need as and when. Would anyone suggest otherwise or alter something? I'm making good progress, I work 2 days a week and thankfully am in a position where I can live comfortably regardless but I'm also 28 and decided this is what I want to do, be a data scientists. I have a top degree in computer science but felt like I learned nothing useful that wasn't already a part of me. I plan to be in full time work by 31 hopefully.
Seriously I hate math very much and had poor performance in my high school and graduation, the terms which you used bounced my mind everytime you said throughout the video . I said oh no.......eventually got motivated by your words. Thanks.
Hi Tina, first of all thank you for the great content you post (I have recently become a subscriber), and I am looking forward to watch this channel grow day by day. I am currently an industrial engineering student, good at math and have learned coding on my own, since the only coding class in my B.S is a Java course. I have even completed the Oracle SQL certificate and am currently looking for a masters degree in Data Science, however I have one major question I would like to have your insight on: Should I start practicing coding on topics apart from data science , maybe learn a bit of software engineering/web development, or should I keep on focusing on my data science skills and practice projects ? IIprogramming/development, so I am kind of confused on how to proceed my journey ?
Coming from a business major, your videos are simple and easy to digest. It easily reinforces my interest in data science and motivates me to pursue it as a career :) Sometimes I feel like my major fall short of the technical skill sets and it’s tough to compete with my peers from computer science background. Any advices for that?
I’m so happy to hear that! Really that is the ultimate purpose of my channel :) if you’re still in school I HIGHLY recommend doing a 5th year master if your college offers it + do a research project with a professor :) that’s the simplest and most direct path IMO to fill in the gap!
i have to admit i am a huge math nerd (and i wanted to go into physics when i was younger), but after going to uni (and not studying physics, may i add) me and math grew apart.... trying to reconnect with it now hhahaha since i'm trying to learn quantitative methods (for international relations, which i studied instead of physics!)
A word of advice I think you should get a pop filter for your mic and then put it a bit closer because I can hear a lot of echoing and a bit of distortion and that's because you're cranking up the gain due to the fact that you're far away from the mic
I recently heard professor Margot Gerritsen in a podcast, she was speaking about how girls are culturally demotivated in the field of mathematics. Subtleties of patriarchy can be seen in meetings in which women opinions are weighted less. With you coming out, smashing this dogma that maths is difficult is making this academia more democratic. I am enjoying data science as in this we can easily combine theory with practice. I wish female percentage in this subject grows from 15% to 50% and we could achieve gender equality practically! :)
That is really really interesting. I added the podcast to my must listen list. Thank you so much for letting me know about it and explaining! I know that math is often discouraged in females but didn’t really know about the subtleties. I really hope so too ❤️❤️
Thank you so much for your video, I'm a life sciences student at U of T, also trying to get into data science/bioinformatics, I was wondering what u of t courses helped you out the most (and courses you wish you took in retrospect).
Ahh that’s a difficult one haha I don’t think anything I took was particularly helpful except the two intro csc classes. They do have a bioinformatics specialist though! Also you should check out oicr at the Mars building and medical biophysics for research. After all uoft is best at research and there’s lots of opportunities in bioinformatics research :)
I’ve never seen math (yet I am actually quite good at math in school) come up in my 3yrs as a data analyst / scientist working for unicorn tech startups... is something seriously wrong?
Hi Tina, nice video. I have a question. Do you think i can learn math for machine learning from an engineering book like advanced engineering mathematics by Erwin Kriszigh?
hey tina. Greetings from Colombia! by chance do you have the possibility of put spanish subs. For me and othe spanish speakers would be of great help. All your content is of high quality and pretty valuable.
@@TinaHuang1 Currently working as a software developer contractor for one of the Canadian Big 5 banks. I transfered out of UofT CS and I'm currently enrolled at Ryerson part time.
Yeah it is! Do you know what I can do to make it better? I’ve played around with the settings a bunch already after watching some tutorials but haven’t been able to make it any better
I'm so good with math, my greatest weakness is programming. Thanks too @kenjee I'm in the #66daysofdata community, it helps me a lot, and thank you too you helped me a lot with SQL, and i hope you do more projects for the future to see how you handle different tasks.
Nice!! 66daysofdata is seriously so awesome and I’m so glad to hear that! Looking forward to the awesome projects you do! And will do - have a couple planned 😊
Math is that subject where you must fill those missing gaps of your knowledge in that chain of mathematics....and after that math will just as easy it is...
@@TinaHuang1 I looked in some of your other videos and figured that only this video has a little bit of an echo. Like you are in a big empty room. Maybe I am just being picky, in that case don't worry about it :)
Oh I see! Hmm I guess I can try to get some carpets haha. Thanks for pointing it out! I’m really trying to up the quality of my videos so feedback really helps ❤️
LOL wait I’m confused because the grade is not bad or it’s bad? Haha regardless I guess I’m comparing it to my other classes because for classes like biology, organic chemistry, psychology etc. I spend like 1/3 the amount of time and end up with a better grade is what I mean 😅