Thanks for watching everyone! Keep up the great work with the #66DaysOfData Challenge! More info on that here: ruclips.net/video/uXLnbdHMf8w/видео.html&ab_channel=KenJee 0:00 Intro 1:30 Why is it important to learn from other's code? 4:04 My approach 5:18 Example
Great video Ken! Learning from other people's code is a great way to learn. Another great way to learn for me is code reviewing where for example some of my colleague or students would send over their code as a Jupyter notebook where I would go over the code, figure out how the ML model workflow could be improved, could be faster, could add some areas of novelty, etc. such that the resulting research paper would have a chance for publication in a scientific journal. Afterwards, I would have discussion with the person who wrote the code and together we can learn from one another on the rationale for choosing a specific approach and whether some component could be adjusted to provide better performance, new areas of novelty, etc. On another note, Amazing video editing (pattern interrupts 😃), also noticed that you're using Notion as placeholders for the code, great tips! (Wow, I've just noticed how long this comment has become)
Thank you. As a beginner , it is okay to just copy other people's notebooks . I didn't know that, and it used to make me miserable . I can make small projects of my own now(really small) , but it still helps to know that back then, I didn't know, what I didn't know.
Yeah, I totally agree with this approach. Whenever I start to work on something, I try to learn the steps form another person's GitHub. If I end with new function or new term I basically Google it out or search in stack. And I found that if I do the same patterns of problems or project for 4/5 times then I will have picked the patterns already. It's amazing, Ken. Thanks for sharing 😊
This is actually so much like machine learning algorithms themselves. Like we go through loads of notebooks to understand the innate algorithms and patterns of being s good data scientist and then after we're done training ourselves with others notebooks, we can test our understanding by predicting on newer projects and then matching with others who have done the similar projects. 🙄
@@KenJee_ds heyy! Firstly, I love your videos and they're a great help and inspiration. But could you like create a basic intro to GitHub video? I searched on youtube but most good ones are for people from traditional comp sci background which I'm not. Thanks!
Until now I used to take on Github notebooks but I think Kaggle is a great option too. Plus there are a lot of introductory notebooks on kaggle if you need to dive in a new concept.
Thanks for explaining the process. Reviewing is the best way which we can apply for learning and I also use discussion like if you have someone with you while studying it leads to good results. I am interacting with you through video but it is really great because it motivates me to come back again and go ahead.
Hi Ken, great work, I am also taking part in the amazing #66DaysOfData, great initiative, it's motivating me so much you have no idea! I can't thank you enough for this. What's the name of the app you're using in the video to take notes? Thanks again
Bro. This is top. Maybe you have already said it and I'm mistaken, but I haven't been able to find it in your videos. What kind of projects would you recommend for getting a data ANALYST job (I've seen the data scientist recommendations, but I suppose they are not the same). I would really appreciate it, although I know you are busy. Thank you for everything ;-)
I think I've talk about it a little in some of the interviews I've done haha. For an analyst, I think you should focus more on visualization projects. A good place to start with that is Tableau and sharing it on Tableau public. I hope this helps!
@@KenJee_ds I'm looking for similar profiles as a beginner, please elaborate on this probably a complete video. Personally, I know basic Tableau, I wanna know where to go with this.
Interestingly, i left beginners project many times just because I could not approach and do them myself. And seeing other's notebook looks like cheating to me. Thank you for removing my notion.
Bro, I am a developer who is thinking to learn but while thinking got stuck much places. Without knowing basis how need to move like that I thought. Basics is completely theoretical. Your way is awesome to learn tq so much❤. I can use this in different field also. I am doing the 66 day challenge even got confusing. You cleared my way which need to proceed. Do well like these videos much. Thank you so much.
So I did a project on linear regression and I'm still kind of struggling to understand it. I looked up Kaggle Notebooks titled "Linear Regression" or something close to that. Other than watch other courses, read other books, and do more projects is there anything else I can do?
I recommend trying quite a few different resources. I would expect one of them will eventually stick. I really like the statquest youtube channel for understanding statistics. I bet he has a linear regression video. You could also try to program it yourself at some point. I find that this is the best way to understand how everything works. It can be very frustrating though!
Always loves your video. You are inspiration I work so hard learning this field. Could you please help me in some of the topics I asked you on #66DaysOfData? I am stucked and I really need some help.
Thanks for the kind words Sameed! I will do my best to answer any youtube comments you post. That is the best place to get a response from me (new comments, not threads I don't get notifications for youtube threads for some reason)
This video was great! I have found some frustration from "not understanding a topic" but not looking for a different explanation. For example, I struggled to understand what a p-value is and how it is useful; to understand what the p-value was, I found a statistics textbook and read about it. I still didn't understand and felt doomed to never understand p-values. Using your approach of looking for more resources, I was able to get a better understanding of it! It is a great idea to see a topic explained in different ways and by different sources. Thank you for bringing this technique up! Thanks for the video Ken!
Great work ...It helps to learn Data science in efficient way ,thanks for this work, by the way can you say the documentation software which you are using it really looks cool
I was doing it all wrong till now. When in earlier video you said go through the notebooks, I was literally reading the code not understanding. Thanks for clearly explaining what we are looking for when we look at the code and how to use the documentations the better way. Thank you so much .
Today is teacher's day in India, so Happy teacher's day Ken. So far your guidance through these videos is really helping me learn Data science. Thank you.
Hi Ken Great content as usual May I have some recommendations or suggestions from you? I'm in my graduate study on modeling fish population dynamics. The situation is that I know about one field, working with data and models in this particular area but I wanna make myself a bit more employable post graduation. Therefore, I start looking up about data science, and realize there are some part of data science work I already do (I just don't know it's called data science out there lol). However, there are some part I haven't touched yet (e.g. computer vision, random forest, and neural network). Should I do machine learning and deep learning courses to organize my knowledge? I deeply understand that nobody can cover all the topics. It's just not happening. What I expect is to be well-prepared for job. Thank you
I think doing some courses is a good idea. I would perhaps focus on random forests and neural networks as they are what a lot of the computer vision stuff is based on. I would also recommend doing projects more so than taking courses. Employers care more about your portfolio than your certificates
@@KenJee_ds can't agree more with the project idea. I just expect doing courses will help me to have a flavor or organizing what I already know, what I don't. I'm doing statistical modelling for my research; hence, there may be many stuff I already know but just don't know what they are called or they are part of DS. Chrs
I've always had this question on how to approach a solved notebook? Should I code open the notebooks side by side and type in one by one or should I go over the solved one and then try to understand. I always wondered how others might be doing or which is the best approach (if any)! So, how would have approached if that was a video lecture where the solution walkthrough is done? Any suggestions, Ken? And, thanks a ton for creating such nuggets of videos, really helpful for beginners like me. Even the comments sections in your videos are insightful.
I think it really depends on how you learn. I usually go through things line by line and run them to see what they do. I usually also delete things or replace things to see what impact that has on the models or graphs. I think the experiment as you go approach is the most useful haha. Thanks for watching my videos Deepa!
I am just a beginner and not an expert, but I think beginners should take the first notebook(by Pedro Marcelino) with a pinch of salt. There are some things I found very wrong in this notebook: 1. You don't always need to drop multicollinear variables. 1 most widely used model in Kaggle is XGBoost and it is immune to multicollinearity. I think, so are other random forest like models. 2. Dropping a variable because it has more than 15% missing data is a bad idea. This threshold is too much in my humble opinion. There are more things which much more experience Data Scientists have pointed out in comments section. I am not saying it's entirely useless but it is not great to follow it 100%. It has maximum votes but maybe it's just an outliar in Kaggle's dataset of Notebooks.
I think this is a totally fair critique! You are 100% right with the missing data. I touch on it a little bit in the video, but I talk about how as you develop more you start to have an understanding of how you personally approach things. When you will often see things that you don't necessarily agree with, and you get to compare it to how you would approach it. I love that you are already at that stage and can think critically about code you see. That is a huge step in the data science learning process!!
As a 20% data scientist 😅 I totally agree with this approach. I also remember you saying anyone who is learning a new thing should seek 30% theoretical knowledge and 70% practice, I also agree with that. Keep up! You are crushing it! 💪
Reviewing the processing code that others have written has exposed a significant issue. Whether the original dataset is already in a "sanitized" state or in need of cleaning, reorganization, or transformation a lot of the existing "notebooks" show there is little to no evidence of getting to know and understand the data before these examples jump to the exploratory data analysis stage. This boils down to blindly accepting the dataset without question. Is the data accurate? What is the source of the underlying data? Perhaps an over reliance on so-called curated or well-known datasets is the underlying cause of this phenomenon. Some people, including instructors, claim the exploratory data analysis stage is when you learn about the dataset. In my opinion and based on experience, the exploratory data analysis stage is supposed to be time evaluate whether the data is suitable for the particular question(s) that instigated the project. Prior to EDA is the time to understand the raw data itself but not whether it can potentially answer those questions. The rush to model building exposes another serious deficiency in the way data science of taught - is the goal of the project to explain historical data or predict future outcomes when presented with new data - is the data skewed or is it relatively normally distributed or should the data be normalized or weighted - is a model necessary? The curated datasets have a place in learning data science. However, too often they instill lemming-like cookie-cutter behaviour. Use these datasets to learn a concept but avoid making them part of a portfolio or treating them as a best practise for real projects. If you want to use them to document your data science learning journey that is fine but at least be creative in the way such datasets are presented - consider demonstrating how a seemingly obvious but incorrect/inappropriate machine learning technique can lead to useless, if not erroneous, conclusions and/or models, and state the reason(s) this technique is inappropriate; consider something other than Python and R to process the data, maybe SQL is adequate; consider integrating other data sources so what-if scenarios can be performed, ideally interactively by end-users. For example, what if data about iceberg flows and the route of the HMS Titanic were added to the model, could the collision have been avoided while still reaching New York City in record time? This might require simulated data based on historical data or modelled using knowledge of ocean currents, fluid dynamics, iceberg and ice fields formation and physical characteristics, weather conditions, etc. Think critically about the data and what if anything it can teach us. Ask the difficult questions and attempt to answer them.
On one hand I really love your channel and I don´t want it to become mainstream but on the other hand I really hope you will be rewarded for the effort & quality that comes with these videos. - So here you get a comment for the algorithm :D Starting my Data Science Bachelor soon & your videos are like steroids for my motiviation :) Keep it up man!
I thought that I hate this field but you help me to realize how much I still love this field and I quit my meaningless and bad job to be better and learn more because that job didn't make me a better data scientist. You even inspire me to write articles. thank you so much.
I really love this initiative! It really motivates me everyday to know that thousands of others are also learning the field! I think reviewing notebooks is a great idea and I'll definitely do that as well in future days-of-data, for now I'll focus on the theory first. Let's build some habits!
So, I have couple of questions related to transformation techniques (log, boxcox etc). - Do I need to use same transformation technique on all the features even though that technique may not be good for other feature? - Do I need to use transformation on dependent variable too? Thanks in advance! again I would like to add that you are doing amazing work!
Thanks for watching! You can transform individual features. Only transform when it is relevant. Sometimes it is practical to transform the dependent variable, but it often isn't needed
Thanks amen! I appreciate you helping others learn faster and more efficiently. I haven’t seen all the resources but it’ll be great you posting more resources to find codes other have built! Looking forward to more!
thanks man! This is helping me unblock things in my head, and allows me to keep my own personal #66daysofdata alive :) Plus, I already wanted to check the very notebook you're showing there (since I am doing a lot of this uncooperative House regression thingy). It also an important lesson since, so far, I wanted to take pride into trying to build my own model alone from scratch without external help ... well, I guess I came as far as I could with this method which your video is helping me understand :)
It is definitely difficult to break into, but it is by no means impossible. Still, I think it is totally ok to not want to put in the amount of work it takes to break in.
Thanks for watching everyone! Keep up the great work with the #66DaysOfData Challenge! More info on that here: ruclips.net/video/uXLnbdHMf8w/видео.html&ab_channel=KenJee
0:00 Intro
1:30 Why is it important to learn from other's code?
4:04 My approach
5:18 Example
Great video Ken! Learning from other people's code is a great way to learn. Another great way to learn for me is code reviewing where for example some of my colleague or students would send over their code as a Jupyter notebook where I would go over the code, figure out how the ML model workflow could be improved, could be faster, could add some areas of novelty, etc. such that the resulting research paper would have a chance for publication in a scientific journal. Afterwards, I would have discussion with the person who wrote the code and together we can learn from one another on the rationale for choosing a specific approach and whether some component could be adjusted to provide better performance, new areas of novelty, etc.
On another note, Amazing video editing (pattern interrupts 😃), also noticed that you're using Notion as placeholders for the code, great tips! (Wow, I've just noticed how long this comment has become)
Haha I like long comments from you! I agree, that is also a great learning approach!
👏 Practical 👏 Tips 👏 FOR THE WIN
💯
Thank you. As a beginner , it is okay to just copy other people's notebooks . I didn't know that, and it used to make me miserable . I can make small projects of my own now(really small) , but it still helps to know that back then, I didn't know, what I didn't know.
Happy this change helped!
May God bless you for sharing this pertinent information. Thank you professor.
Thanks for watching Sirak!
Yeah, I totally agree with this approach. Whenever I start to work on something, I try to learn the steps form another person's GitHub. If I end with new function or new term I basically Google it out or search in stack. And I found that if I do the same patterns of problems or project for 4/5 times then I will have picked the patterns already.
It's amazing, Ken.
Thanks for sharing 😊
Glad to hear you're experimenting with this! Thanks for watching Thinam!
Ken, Thanks for the great insight. Very useful video for beginners.
Glad to hear you found it useful! This concept was completely new to me when I started
This is actually so much like machine learning algorithms themselves. Like we go through loads of notebooks to understand the innate algorithms and patterns of being s good data scientist and then after we're done training ourselves with others notebooks, we can test our understanding by predicting on newer projects and then matching with others who have done the similar projects. 🙄
Love this analogy!
@@KenJee_ds heyy! Firstly, I love your videos and they're a great help and inspiration. But could you like create a basic intro to GitHub video? I searched on youtube but most good ones are for people from traditional comp sci background which I'm not. Thanks!
@@sukantasaha5678 Haha, honestly I need to up my github game myself. Will make one after I review it!
Until now I used to take on Github notebooks but I think Kaggle is a great option too. Plus there are a lot of introductory notebooks on kaggle if you need to dive in a new concept.
Kaggle is definitely a great option!
Thank you so much, Ken! You are truly amazing!!!
Thank you for watching Jing!!
Thanks for explaining the process. Reviewing is the best way which we can apply for learning and I also use discussion like if you have someone with you while studying it leads to good results. I am interacting with you through video but it is really great because it motivates me to come back again and go ahead.
Glad to hear it is motivating! This was one of the biggest changes in my own personal learning!
@@KenJee_ds That's great!!!
Great process! I should look into this Notion app.
Hi Ken, great work, I am also taking part in the amazing #66DaysOfData, great initiative, it's motivating me so much you have no idea! I can't thank you enough for this. What's the name of the app you're using in the video to take notes? Thanks again
Thanks for following along. The app is called Notion!
Bro. This is top. Maybe you have already said it and I'm mistaken, but I haven't been able to find it in your videos. What kind of projects would you recommend for getting a data ANALYST job (I've seen the data scientist recommendations, but I suppose they are not the same). I would really appreciate it, although I know you are busy. Thank you for everything ;-)
I think I've talk about it a little in some of the interviews I've done haha. For an analyst, I think you should focus more on visualization projects. A good place to start with that is Tableau and sharing it on Tableau public. I hope this helps!
Thank you so much ♥️
@@KenJee_ds I'm looking for similar profiles as a beginner, please elaborate on this probably a complete video. Personally, I know basic Tableau, I wanna know where to go with this.
Interestingly, i left beginners project many times just because I could not approach and do them myself. And seeing other's notebook looks like cheating to me. Thank you for removing my notion.
I promise it's not cheating haha!
Bro, I am a developer who is thinking to learn but while thinking got stuck much places. Without knowing basis how need to move like that I thought. Basics is completely theoretical. Your way is awesome to learn tq so much❤. I can use this in different field also. I am doing the 66 day challenge even got confusing. You cleared my way which need to proceed. Do well like these videos much. Thank you so much.
Glad to hear this cleared up your way to proceed! Thanks for watching!
So I did a project on linear regression and I'm still kind of struggling to understand it. I looked up Kaggle Notebooks titled "Linear Regression" or something close to that. Other than watch other courses, read other books, and do more projects is there anything else I can do?
I recommend trying quite a few different resources. I would expect one of them will eventually stick. I really like the statquest youtube channel for understanding statistics. I bet he has a linear regression video. You could also try to program it yourself at some point. I find that this is the best way to understand how everything works. It can be very frustrating though!
Thanks for valuable tips Ken. I am going to apply them on my #66DaysOfData journey and projects from now on
Thanks for watching! Love that you're going to apply them!!
Always loves your video. You are inspiration I work so hard learning this field. Could you please help me in some of the topics I asked you on #66DaysOfData? I am stucked and I really need some help.
Thanks for the kind words Sameed! I will do my best to answer any youtube comments you post. That is the best place to get a response from me (new comments, not threads I don't get notifications for youtube threads for some reason)
Amazing video Ken! This is so insightful and helpful for my data science journey
Yes! That's what I love to hear. Glad it was helpful!
This video was great!
I have found some frustration from "not understanding a topic" but not looking for a different explanation. For example, I struggled to understand what a p-value is and how it is useful; to understand what the p-value was, I found a statistics textbook and read about it. I still didn't understand and felt doomed to never understand p-values. Using your approach of looking for more resources, I was able to get a better understanding of it! It is a great idea to see a topic explained in different ways and by different sources. Thank you for bringing this technique up!
Thanks for the video Ken!
Glad some experimentation helped you here! It was something that changed the game for me!
Great work ...It helps to learn Data science in efficient way ,thanks for this work,
by the way can you say the documentation software which you are using it really looks cool
It is called the notion app! Thanks for watching!
I was doing it all wrong till now. When in earlier video you said go through the notebooks, I was literally reading the code not understanding. Thanks for clearly explaining what we are looking for when we look at the code and how to use the documentations the better way. Thank you so much .
Glad to hear this helped clear some things up Piyush!
Which tool is he using for the notes again!?
The Notion app!
Thank u sir for this video You clear all my doubts.🙂🙂🙂
Yes! Glad to hear it helped clear some things up!
Thank you Ken. I'll adopt this approach the more.
Awesome!
Thanks Ken. Got a clear idea of how to go about learning data science. Following along with you on the #66daysofdata challenge.
Thanks for following along!
Today is teacher's day in India, so Happy teacher's day Ken. So far your guidance through these videos is really helping me learn Data science. Thank you.
Such a cool tradition. Thank you for being a loyal subscriber!
Happy Teacher's Day ken, Thanks for sharing...🧡
Thank you!!
Great video Ken! Your daily progress updates on discord motivates me even more. Thank you for doing this😊
Glad to hear!! Doing my best to stay motivated with everyone!
Hi Ken
Great content as usual
May I have some recommendations or suggestions from you?
I'm in my graduate study on modeling fish population dynamics. The situation is that I know about one field, working with data and models in this particular area but I wanna make myself a bit more employable post graduation. Therefore, I start looking up about data science, and realize there are some part of data science work I already do (I just don't know it's called data science out there lol). However, there are some part I haven't touched yet (e.g. computer vision, random forest, and neural network).
Should I do machine learning and deep learning courses to organize my knowledge?
I deeply understand that nobody can cover all the topics. It's just not happening. What I expect is to be well-prepared for job.
Thank you
I think doing some courses is a good idea. I would perhaps focus on random forests and neural networks as they are what a lot of the computer vision stuff is based on. I would also recommend doing projects more so than taking courses. Employers care more about your portfolio than your certificates
@@KenJee_ds can't agree more with the project idea.
I just expect doing courses will help me to have a flavor or organizing what I already know, what I don't. I'm doing statistical modelling for my research; hence, there may be many stuff I already know but just don't know what they are called or they are part of DS.
Chrs
This is a great approach for people who don't have a lot of free time on their hands. Another amazing video Ken!
Thanks for watching, I agree!
I've always had this question on how to approach a solved notebook? Should I code open the notebooks side by side and type in one by one or should I go over the solved one and then try to understand. I always wondered how others might be doing or which is the best approach (if any)! So, how would have approached if that was a video lecture where the solution walkthrough is done? Any suggestions, Ken? And, thanks a ton for creating such nuggets of videos, really helpful for beginners like me. Even the comments sections in your videos are insightful.
I think it really depends on how you learn. I usually go through things line by line and run them to see what they do. I usually also delete things or replace things to see what impact that has on the models or graphs. I think the experiment as you go approach is the most useful haha. Thanks for watching my videos Deepa!
@@KenJee_ds Thanks Ken for the response.
I am just a beginner and not an expert, but I think beginners should take the first notebook(by Pedro Marcelino) with a pinch of salt. There are some things I found very wrong in this notebook:
1. You don't always need to drop multicollinear variables. 1 most widely used model in Kaggle is XGBoost and it is immune to multicollinearity. I think, so are other random forest like models.
2. Dropping a variable because it has more than 15% missing data is a bad idea. This threshold is too much in my humble opinion.
There are more things which much more experience Data Scientists have pointed out in comments section.
I am not saying it's entirely useless but it is not great to follow it 100%. It has maximum votes but maybe it's just an outliar in Kaggle's dataset of Notebooks.
I think this is a totally fair critique! You are 100% right with the missing data. I touch on it a little bit in the video, but I talk about how as you develop more you start to have an understanding of how you personally approach things. When you will often see things that you don't necessarily agree with, and you get to compare it to how you would approach it. I love that you are already at that stage and can think critically about code you see. That is a huge step in the data science learning process!!
As a 20% data scientist 😅 I totally agree with this approach. I also remember you saying anyone who is learning a new thing should seek 30% theoretical knowledge and 70% practice, I also agree with that.
Keep up! You are crushing it!
💪
Glad you agree! Thanks for watching and for the kind words!
one of the best data science channels out there.. kia kaha tonu..keep going
Thank you for the kind words! Doing my best to keep making useful content for you!
Reviewing the processing code that others have written has exposed a significant issue. Whether the original dataset is already in a "sanitized" state or in need of cleaning, reorganization, or transformation a lot of the existing "notebooks" show there is little to no evidence of getting to know and understand the data before these examples jump to the exploratory data analysis stage. This boils down to blindly accepting the dataset without question. Is the data accurate? What is the source of the underlying data?
Perhaps an over reliance on so-called curated or well-known datasets is the underlying cause of this phenomenon. Some people, including instructors, claim the exploratory data analysis stage is when you learn about the dataset. In my opinion and based on experience, the exploratory data analysis stage is supposed to be time evaluate whether the data is suitable for the particular question(s) that instigated the project. Prior to EDA is the time to understand the raw data itself but not whether it can potentially answer those questions. The rush to model building exposes another serious deficiency in the way data science of taught - is the goal of the project to explain historical data or predict future outcomes when presented with new data - is the data skewed or is it relatively normally distributed or should the data be normalized or weighted - is a model necessary?
The curated datasets have a place in learning data science. However, too often they instill lemming-like cookie-cutter behaviour. Use these datasets to learn a concept but avoid making them part of a portfolio or treating them as a best practise for real projects. If you want to use them to document your data science learning journey that is fine but at least be creative in the way such datasets are presented - consider demonstrating how a seemingly obvious but incorrect/inappropriate machine learning technique can lead to useless, if not erroneous, conclusions and/or models, and state the reason(s) this technique is inappropriate; consider something other than Python and R to process the data, maybe SQL is adequate; consider integrating other data sources so what-if scenarios can be performed, ideally interactively by end-users. For example, what if data about iceberg flows and the route of the HMS Titanic were added to the model, could the collision have been avoided while still reaching New York City in record time? This might require simulated data based on historical data or modelled using knowledge of ocean currents, fluid dynamics, iceberg and ice fields formation and physical characteristics, weather conditions, etc.
Think critically about the data and what if anything it can teach us. Ask the difficult questions and attempt to answer them.
Very valuable tips!!! Thank you so much for taking time to share!
On one hand I really love your channel and I don´t want it to become mainstream but on the other hand I really hope you will be rewarded for the effort & quality that comes with these videos. - So here you get a comment for the algorithm :D
Starting my Data Science Bachelor soon & your videos are like steroids for my motiviation :) Keep it up man!
I appreciate it Jonas! I will do my best to keep my style and connection with subscribers like yourself going! Good luck in your degree!!
Thanks KEN keep up the good work!
Thanks for watching!!
I thought that I hate this field but you help me to realize how much I still love this field and I quit my meaningless and bad job to be better and learn more because that job didn't make me a better data scientist. You even inspire me to write articles. thank you so much.
Glad I have help you find your love!!
Thanks ken for being you :)
Haha I don't know how to be any other way! Thanks for being such a loyal subscriber!
You are such an inspiration to us!!
Thanks for motivating us :)
Thanks for watching!!
wow, thanks for sharing.
Thanks for watching!
This was super helpful ! :)
Thanks for watching!
always the best tips, thx man
Thanks for watching Gabriel! Glad you liked them!
I really love this initiative! It really motivates me everyday to know that thousands of others are also learning the field!
I think reviewing notebooks is a great idea and I'll definitely do that as well in future days-of-data, for now I'll focus on the theory first.
Let's build some habits!
Glad to hear it keeps you motivated Boris! Your comments keep me motivated!
I'm glad!!
Let's help ken reach 💯 k 🔥
😃
zero dislikes wow!!!
Haha We all get lucky sometimes!
can I know which is the note taking app which you are using?
It is called Notion!
@@KenJee_ds thank you so much can u share ur template please
You are a life saver !!
Thanks for watching!
So, I have couple of questions related to transformation techniques (log, boxcox etc).
- Do I need to use same transformation technique on all the features even though that technique may not be good for other feature?
- Do I need to use transformation on dependent variable too?
Thanks in advance! again I would like to add that you are doing amazing work!
Thanks for watching! You can transform individual features. Only transform when it is relevant. Sometimes it is practical to transform the dependent variable, but it often isn't needed
Thanks amen! I appreciate you helping others learn faster and more efficiently. I haven’t seen all the resources but it’ll be great you posting more resources to find codes other have built! Looking forward to more!
Doing my best to post more resources!! Thanks for watching!
thanks man! This is helping me unblock things in my head, and allows me to keep my own personal #66daysofdata alive :) Plus, I already wanted to check the very notebook you're showing there (since I am doing a lot of this uncooperative House regression thingy). It also an important lesson since, so far, I wanted to take pride into trying to build my own model alone from scratch without external help ... well, I guess I came as far as I could with this method which your video is helping me understand :)
Great stuff!!! Early on we need to put our pride to the side and just absorb. You will be able to build your own great models eventually!
i feel im gonna quit this field the barrier to entry is too high and i cant build my project out of novel ideas
It is definitely difficult to break into, but it is by no means impossible. Still, I think it is totally ok to not want to put in the amount of work it takes to break in.
Wow. Best video I’ve seen for learning date science / analysis
Thanks for the kind words. Glad you found it helpful!
This is a highly valuable and useful video!
Thanks a ton Ken!🙌
Thanks for watching Mohak!
Thank you, you answered the exact question I had, how to best go about this.
Yes! I love it when videos do that haha. Thank you for watching!
This is dope bro! Seeing you through the process is priceless. Good job. ty
Thanks for watching!!
Exactly what I was looking for
Thank you man
Yes! Thanks for watching!
Just curious, but why do you have 3 clocks in your room😅
I actually have 4 haha. My business operates in many different time zones, so I use them so I don't miss my meetings
Love your content!
Willing to start the #66DaysofData
Kudos!!!!!!!!
Thank you! Looking forward to having you on board!
I love hands-on learning
Glad you love it!! It is the best way to do it IMOP
This will be what I do
Great!!
Genius
Haha I don't know about that, but I'm glad you found it helpful!
Happy Teacher's Day!!!
Thank you!
I think this is actually an incredible idea
Thanks Javier!! Would love to hear how it works for you!
Excellent tips 👍
Thanks for watching!!
ofcos he didnt look good
I need some excuse haha