Thanks for watching everyone! I'm having a lot of fun with the intros, I hope you all like them :). If you don't, I added chapters at the bottom so you can skip it easily! Watch these next: 5 essential data science projects: th-cam.com/video/BBDiadC8BvE/w-d-xo.html&ab_channel=KenJee Titanic Tutorial: th-cam.com/video/I3FBJdiExcg/w-d-xo.html&ab_channel=KenJee #66DaysOfData: th-cam.com/video/uXLnbdHMf8w/w-d-xo.html&ab_channel=KenJee Data Science Project from Scratch: th-cam.com/play/PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t.html
Hello, i have been into DS journey for 4 months now; i learned pretty much all python, learned sql(didnt use much though), i learned basic ML so i know linear/logistic regression, decision tree, random forests. Also went to Kaggle and did some courses and still doing. Meanwhile i decided to try doing some tests on data and i found that i really struggled. It took me 2 weeks just on the titanic dataset because i felt many times that i dont know this or that and had to look up documentation and visit stackoverflow . Long story short i got to try the few estimators i know which made me see how little i know, i havent felt dejected or depressed to give up. But this feeling of not knowing enough is bothering me, also since i cant really pay for any of the courses online the free kaggle stuff seems to be not enough. I feel a little lost. Maybe this feels like ranting or a newbie's mistakes but can u or anyobody give some advice? Im on the stage where i know what i dont know, now i need a way to step by step learn all that. Freecodecamp maybe? I remember you said look at other people's code, i will do that.
Spending a bit of time to find a good idea definitely is worth it! Once you find something you are truly passionate about, the motivation can sometimes be limitless. In my opinion, one great project is worth more than a few okay ones, because the passion can really shine through. (Also, awesome intro :D)
Thank you for this. I'm tired of spending time watching incredibly detailed ang long video tutorials about data science without actually applying what I've learned to a project.
"#1. Uber for dogs" - my business name submission is "Duber" 😂 Great video! I loved the creative and comical aspects you add into your videos. Solid advice!!
Wow, the video editing is amazing Ken! Lot's of nicely placed puns and humor. I totally agree with your advice that coming up with project ideas is a great skill to have. In academia, one of the most important skills to be acquired during graduate study is essentially this skill of coming up with project ideas (finding solutions for solving biological data problems). Thanks for breaking down this important topic in a compact and entertaining video 😆👍
My first ever comment on YT (you are that good :) ) : thank you for these techniques. For total newbies like I am, that are not looking (yet) for originality but just build / practice basic skills (and self confidence in data science) few concrete common projects would help a lot. I get that it might not be super exciting for experts but your walk through Titanic was great. Maybe few more like these? (Disclaimer: I am really at the beginning , did not even start the Titanic project yet)
Ken to the rescue once again! I'm currently in the process of planning what my next project should be haha! I see the new merchandise is up for sale now too 👀
Could you make a video on all the skills you have currently in regards to data science. it could be like reding your CV in aspect of data science. is just python and tablue enough to start applying for this role. Have completed your Model_Building series it was a awesome pathway towards DS
You're really starting to get the hang of this TH-cam(r) business down! The editing is looking great with the different camera angles and transitions and stuff. Have you considered streaming on Twitch or TH-cam for some extra views/content? Resume reviews, portfolio reviews, interview advice/practice, teaching something, etc? I know you put out that course a while ago. Good content in the video by the way. I've been using Kaggle to learn some proper ML techniques and their implementations along with when/how to use them, and also get acquainted with Python more since I'm an such an R junkie. Playing with some stock data sets on there has been good for me, but I do agree, value or good business value comes from gathering some original data or doing something with some impact. I'm currently thinking about trying to build a special app/thingy (currently thinking an RShiny web app) that scrapes data/information on Yugioh's card collecting/valuation website and then so someone can enter what cards they have and how much (ignoring PSA value) and then it'll autopopulate the current valuation cell next to it or something. One small note, and I hope this doesn't come across as offensive, but I personally thought that the camera angle/cutaway/transition at around @6:45 was a bit awkward, as it looked like you were speaking but there was an audio lag or something. Very well could just be something due to my internet, but it took me out a little bit if you know what I mean. Nonetheless though, great video and great content. This channel has been immensely helpful and inspiring to me. Keep up the good work! Hopefully this large comment adds valuable data to your corpus you're building. Sorry for the special characters you'll have to get rid of when you tokenize and clean the data lol.
Thanks for the kind words and constructive feedback Matthew! I will be doing a live stream next week for 100k subscribers. They are ok, but I personally like creating something a bit more polished than just putting it all out there. The project sounds awesome! Would love to see it when you're done! No offense taken, I agree with you about the 6:55 mark, will be learning from that one lol!
Great tips ken but I have a really dumb question xD. Suppose I have a great project idea (lets say agriculture related), but now from where will I get the dataset?
Good info as always, definitely liked the analogy to coming up with new business ideas and being more specific. Question, in the sports analytics industry how much of your time is spent on various activities? Maintaining and updating the current data analytics? Answering new questions that are brought up by your internal customers? Answering your own questions or coming up with new data analytics aproaches to add value?
It really varies by the position, but I spend about 1/3 of my time managing my team, 1/3 working on existing work, and 1/3 interacting with clients and gathering requirements. I hope this helps!
@@KenJee_ds Thanks for the feedback, that was exactly what I was looking for. Do you spend any time looking for or researching something new or novel to try to provide value not based specifically on their requirements?
@@MrZola1234 I split the time interacting and gathering requirements, getting their feedback but also coming up with our team's own solutions. I hope this helps!
WHAT?!! Good projects cannot be found in your closet, laundry, desk drawers etc ... WHY AM I ALWAYS THE LAST TO KNOW?!!!! More seriously, right now, it's not so much ideas that I have problems with at the moment, it's finding enough data to run them ... yep, that time is closing faster than I thought where I will have to learn to scrape data for myself! Especially since many of the ideas I have are around using specific type of pictures (loads of them fantasy, science fiction .... also had an idea which will requires 'solar punk' pictures, all of that in the name of ART and by using ML :o) ). But, more likely, it will be about scrapping for text, especially released scripts from some of my favorite series ... or old books!! ( I wonder if I can do a 'whosdunnit' NLP thingy :o) ) ... so many choices, but doesn't choice imply renouncing some of said choices ? ;o)
In my opinion, you should stick to the field you are interested in. And data should be complicated enough to be worth analyzing, not like two columns. But to get both the data and the idea is the biggest problem for me so far. I either have the idea, but it's nearly impossible to find any decent amount of data or there would be lots of data, but on a topic that I've never found the least exciting. Sometimes data isn't complicated enough for a project. For instance, I do monitor used cars sales on a regular basis, I just love reading about cars, comparing them. But the tricky thing is, there are like 5 or 6 variables to utilize, like mileage, color, horsepower, year produced etc. It's more than enough for eda, but there is not much room for machine learning project. The only interesting idea I came up with (apart fromobvious linear regression) is to analyze, whether mileage is real or not based on the price, number of owners etc. It seems like some fields are inferior in terms of generating data, than others
You're right, collecting data can definitely be a huge bottleneck. I also think that even if you don't have a lot of data, you can change the nature of the product. You can show off your productionization skills by make a simple web app or tool instead of just doing EDA. Thanks again for watching!
Hi Kenji. Is there a trade-off between projects that are personally interesting vs portfoilo-worthy projects? Do your projects have to be directly tailored to whatever industry you're trying to get into or is there such a thing as a general data science portfolio? Thanks for your videos.
I don't think there is. Most employers will be looking for you to do something unique with data. I find that personally interesting projects are some of the most unique you can do. There is no such thing as a general data science project portfolio in terms of which data you should use. You do generally want to showcase that you can do classification, clustering, regression, and some deep learning through your projects though.
Thank you so much! i have a question that's been bothering me, if someone worked on your insightful jobs salary project, like , started it from scratch and made changes ( different eda/including ann model..) through out the project and learned a lot from it, could they include it in their resume ? if yes, i think we must reference the owner of the idea ? or is it just better to not include it ? your project inspired me a lot and made me love data science more, thank you so much and hope you will answer my question :)
Definitely ok to include it in your resume! You should also cite the other project and the parts of it that you were influenced by. Glad to hear the project inspired you :)
@@KenJee_ds it took me 2 months to go through all the project, because every single time you mentioned a new concept , i stopped and went to learn about it and then came back to the project, Yes i will include it with many others, thank you so much for all the work you've put out for us on your platform !
Hey ken I am trying to do a project on the past NBA season. Do you have any ideas where I could find extensive nba data(besides stats.nba.com)? Thanks!!
Definitely not trying to say that! Building models is important. What I'm trying to say is that the complexity of models tat you use is less important than the outcome of the project for the stakeholder. I hope this clears things up!
@@KenJee_ds Yes, thank you so much. Final question: If an employer tells you that you will not build any model, can this job be named as "Data Scientist"?
Working on it! I should have one coming out at the end of the month where I analyze my youtube comments. The projects take me a bit longer, and my real work has kind of been getting the better of me :(
I don't think I could manage multiple channels outside of the Discord. I recommend exploring that though! I will consider it if slack gets enough support!
How afraid should we be about choosing projects that end up failing? Sometimes I get caught in this idea that I need to find the perfect projects that I know would be successful before I even start exploring the data.
Don't be afraid! Even if the project outcome isn't exactly what you wanted, at least you learned something. Honestly, I don't think I've ever had a project go as I planned. What is important is that you adjust as you go and do your best to find something interesting! I hope this helps!
It would be extremely difficult, but I think it would be possible. To get in without a degree, you would have to do something extremely impressive on kaggle or in your own personal projects.
![[LIVE] : ONE ลุมพินี 79 | คู่เอก "ก้องชัย vs อาเมียร์"](http://i.ytimg.com/vi/WfPSA1A0NHc/mqdefault.jpg)
Thanks for watching everyone! I'm having a lot of fun with the intros, I hope you all like them :). If you don't, I added chapters at the bottom so you can skip it easily!
Watch these next:
5 essential data science projects: th-cam.com/video/BBDiadC8BvE/w-d-xo.html&ab_channel=KenJee
Titanic Tutorial: th-cam.com/video/I3FBJdiExcg/w-d-xo.html&ab_channel=KenJee
#66DaysOfData: th-cam.com/video/uXLnbdHMf8w/w-d-xo.html&ab_channel=KenJee
Data Science Project from Scratch: th-cam.com/play/PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t.html
Hello, i have been into DS journey for 4 months now; i learned pretty much all python, learned sql(didnt use much though), i learned basic ML so i know linear/logistic regression, decision tree, random forests. Also went to Kaggle and did some courses and still doing. Meanwhile i decided to try doing some tests on data and i found that i really struggled. It took me 2 weeks just on the titanic dataset because i felt many times that i dont know this or that and had to look up documentation and visit stackoverflow . Long story short i got to try the few estimators i know which made me see how little i know, i havent felt dejected or depressed to give up. But this feeling of not knowing enough is bothering me, also since i cant really pay for any of the courses online the free kaggle stuff seems to be not enough. I feel a little lost.
Maybe this feels like ranting or a newbie's mistakes but can u or anyobody give some advice?
Im on the stage where i know what i dont know, now i need a way to step by step learn all that. Freecodecamp maybe? I remember you said look at other people's code, i will do that.
I came to know your channel when you uploaded the video "How i would do data science (if i had to start over)" . 🤗😁
Spending a bit of time to find a good idea definitely is worth it!
Once you find something you are truly passionate about, the motivation can sometimes be limitless.
In my opinion, one great project is worth more than a few okay ones, because the passion can really shine through.
(Also, awesome intro :D)
Thanks Boris!!!
Thank you for this. I'm tired of spending time watching incredibly detailed ang long video tutorials about data science without actually applying what I've learned to a project.
Thanks for watching Martin! Glad you found the shorter videos helpful!
"#1. Uber for dogs" - my business name submission is "Duber"
😂
Great video! I loved the creative and comical aspects you add into your videos. Solid advice!!
Incredible name! I should definitely start that one haha. I was hoping someone would read the awful ideas I put down :)
Thanks for watching!!!
@@KenJee_ds Little lizard hats was a close second. Solid ideas! Hahah thanks for the laugh today
@@ashleyhaa :)
Wow, the video editing is amazing Ken! Lot's of nicely placed puns and humor. I totally agree with your advice that coming up with project ideas is a great skill to have. In academia, one of the most important skills to be acquired during graduate study is essentially this skill of coming up with project ideas (finding solutions for solving biological data problems). Thanks for breaking down this important topic in a compact and entertaining video 😆👍
Thanks data professor! Doing my best to integrate more of my real personality in my videos haha
My first ever comment on YT (you are that good :) ) : thank you for these techniques. For total newbies like I am, that are not looking (yet) for originality but just build / practice basic skills (and self confidence in data science) few concrete common projects would help a lot. I get that it might not be super exciting for experts but your walk through Titanic was great. Maybe few more like these? (Disclaimer: I am really at the beginning , did not even start the Titanic project yet)
Thanks for the comment!!! I am hoping to do more beginner projects before the beginning of the year :)
Ken to the rescue once again! I'm currently in the process of planning what my next project should be haha!
I see the new merchandise is up for sale now too 👀
Glad to hear the video was helpful Mario! Hopefully my test batch of shirts gets in soon!
Great channel dude! thank for share your work.
Best regards from Chile
Thanks!!
Amazing editing!! And super solid advice! :D
Thanks Tina!!!
Great advice, thanks!
Thanks for watching Neel!
These 3 techniques will really make an impact on me to go ahead. Problem identification is a very much important skill in data science.
Glad to hear they were impactful! Also, please send me an email at kenjee.ds@gmail.com! You were one of the winners of the leaderboard for January!
@@KenJee_ds Yeah... I will and Thank you...
Master Master Ken, this video was great, the quality of your videos is going up. Thanks bro!
Thanks Miguel!!!
Thank you! ☺ You are like a project guide I also searched for during my Bachelor's degree!!!!!! 😭😭
Glad I could help!
Ooohh, this video is on next level ken :) Amazing techniques too!
Thanks Jithendra!! Glad you liked it :)
Could you make a video on all the skills you have currently in regards to data science. it could be like reding your CV in aspect of data science. is just python and tablue enough to start applying for this role. Have completed your Model_Building series it was a awesome pathway towards DS
I will look into that! Seems like a good topic!
You're really starting to get the hang of this TH-cam(r) business down! The editing is looking great with the different camera angles and transitions and stuff. Have you considered streaming on Twitch or TH-cam for some extra views/content? Resume reviews, portfolio reviews, interview advice/practice, teaching something, etc? I know you put out that course a while ago. Good content in the video by the way. I've been using Kaggle to learn some proper ML techniques and their implementations along with when/how to use them, and also get acquainted with Python more since I'm an such an R junkie. Playing with some stock data sets on there has been good for me, but I do agree, value or good business value comes from gathering some original data or doing something with some impact.
I'm currently thinking about trying to build a special app/thingy (currently thinking an RShiny web app) that scrapes data/information on Yugioh's card collecting/valuation website and then so someone can enter what cards they have and how much (ignoring PSA value) and then it'll autopopulate the current valuation cell next to it or something. One small note, and I hope this doesn't come across as offensive, but I personally thought that the camera angle/cutaway/transition at around @6:45 was a bit awkward, as it looked like you were speaking but there was an audio lag or something. Very well could just be something due to my internet, but it took me out a little bit if you know what I mean. Nonetheless though, great video and great content. This channel has been immensely helpful and inspiring to me. Keep up the good work! Hopefully this large comment adds valuable data to your corpus you're building. Sorry for the special characters you'll have to get rid of when you tokenize and clean the data lol.
Thanks for the kind words and constructive feedback Matthew! I will be doing a live stream next week for 100k subscribers. They are ok, but I personally like creating something a bit more polished than just putting it all out there.
The project sounds awesome! Would love to see it when you're done!
No offense taken, I agree with you about the 6:55 mark, will be learning from that one lol!
Going to be amazing seeing the title but I will go and watch the video now💯
I hope you enjoyed it Shruti!
@@KenJee_ds yes I did very informative and useful ❤️💯
@@shrutijain1628 Great!!
Great tips ken but I have a really dumb question xD. Suppose I have a great project idea (lets say agriculture related), but now from where will I get the dataset?
There are probably a lot of good agriculture datasets on kaggle. If the data doesn't exist there, you could try scraping it too!
@@KenJee_ds alright!
Congrats Ken, You have achieved 100k victory successfuly
Thanks Muhammad!
Good info as always, definitely liked the analogy to coming up with new business ideas and being more specific.
Question, in the sports analytics industry how much of your time is spent on various activities? Maintaining and updating the current data analytics? Answering new questions that are brought up by your internal customers? Answering your own questions or coming up with new data analytics aproaches to add value?
It really varies by the position, but I spend about 1/3 of my time managing my team, 1/3 working on existing work, and 1/3 interacting with clients and gathering requirements. I hope this helps!
@@KenJee_ds Thanks for the feedback, that was exactly what I was looking for. Do you spend any time looking for or researching something new or novel to try to provide value not based specifically on their requirements?
@@MrZola1234 I split the time interacting and gathering requirements, getting their feedback but also coming up with our team's own solutions. I hope this helps!
heyy nice video! i have a question, what technique do you use to scrapp data from the comment section of your videos??? i love your channel by the way
Thanks for watching! I used the TH-cam API!
WHAT?!! Good projects cannot be found in your closet, laundry, desk drawers etc ... WHY AM I ALWAYS THE LAST TO KNOW?!!!! More seriously, right now, it's not so much ideas that I have problems with at the moment, it's finding enough data to run them ... yep, that time is closing faster than I thought where I will have to learn to scrape data for myself! Especially since many of the ideas I have are around using specific type of pictures (loads of them fantasy, science fiction .... also had an idea which will requires 'solar punk' pictures, all of that in the name of ART and by using ML :o) ).
But, more likely, it will be about scrapping for text, especially released scripts from some of my favorite series ... or old books!! ( I wonder if I can do a 'whosdunnit' NLP thingy :o) ) ... so many choices, but doesn't choice imply renouncing some of said choices ? ;o)
Informative Video Ken.
Thanks for watching Mohit!!
Oo the editing is getting next level 🔥
Thanks Nash! I'm having a little bit too much fun with it now haha
@@KenJee_ds The B-Roll when you said exploratory was elite
In my opinion, you should stick to the field you are interested in. And data should be complicated enough to be worth analyzing, not like two columns. But to get both the data and the idea is the biggest problem for me so far. I either have the idea, but it's nearly impossible to find any decent amount of data or there would be lots of data, but on a topic that I've never found the least exciting. Sometimes data isn't complicated enough for a project. For instance, I do monitor used cars sales on a regular basis, I just love reading about cars, comparing them. But the tricky thing is, there are like 5 or 6 variables to utilize, like mileage, color, horsepower, year produced etc. It's more than enough for eda, but there is not much room for machine learning project. The only interesting idea I came up with (apart fromobvious linear regression) is to analyze, whether mileage is real or not based on the price, number of owners etc. It seems like some fields are inferior in terms of generating data, than others
You're right, collecting data can definitely be a huge bottleneck. I also think that even if you don't have a lot of data, you can change the nature of the product. You can show off your productionization skills by make a simple web app or tool instead of just doing EDA. Thanks again for watching!
Hi Kenji. Is there a trade-off between projects that are personally interesting vs portfoilo-worthy projects? Do your projects have to be directly tailored to whatever industry you're trying to get into or is there such a thing as a general data science portfolio? Thanks for your videos.
I don't think there is. Most employers will be looking for you to do something unique with data. I find that personally interesting projects are some of the most unique you can do. There is no such thing as a general data science project portfolio in terms of which data you should use. You do generally want to showcase that you can do classification, clustering, regression, and some deep learning through your projects though.
smashed the like button...😉👍
Thanks haha!
Thank you so much! i have a question that's been bothering me, if someone worked on your insightful jobs salary project, like , started it from scratch and made changes ( different eda/including ann model..) through out the project and learned a lot from it, could they include it in their resume ? if yes, i think we must reference the owner of the idea ? or is it just better to not include it ? your project inspired me a lot and made me love data science more, thank you so much and hope you will answer my question :)
Definitely ok to include it in your resume! You should also cite the other project and the parts of it that you were influenced by. Glad to hear the project inspired you :)
@@KenJee_ds it took me 2 months to go through all the project, because every single time you mentioned a new concept , i stopped and went to learn about it and then came back to the project, Yes i will include it with many others, thank you so much for all the work you've put out for us on your platform !
@@ogawasan5581 Thank you for following along! I think you took a really good approach to learning the project!
Ken is back at it again with his papayas haha
Flaunt it if you got it!
Why do TH-camrs not edit out the part of the shot where they are reaching for the camera off-button? Is it a meme?
It is a laziness lol
Looks like you are having waaaaay too much fun filming the b rolls 😂
Wayyyy too much fun lol
Hey ken I am trying to do a project on the past NBA season. Do you have any ideas where I could find extensive nba data(besides stats.nba.com)? Thanks!!
I would try Kaggle.com! They have a surprising amount of sports data!
Gave you a like for the valorant name drop
hahaha
7:48 Are you trying to say that building models is not necessary? Could you explain your argument here?
Definitely not trying to say that! Building models is important. What I'm trying to say is that the complexity of models tat you use is less important than the outcome of the project for the stakeholder. I hope this clears things up!
@@KenJee_ds Yes, thank you so much. Final question: If an employer tells you that you will not build any model, can this job be named as "Data Scientist"?
@@halildurmaz7827 It could still be called that techincally
Problem found: The "subscribe" animation still says 64k followers, let's fix that! :D thx for the video
Haha working on it!
02:40 What were your 10 ideas to improve airline flights? Just curious.
May have to get back to you on that one haha. I have no clue where the notebook with that one is...
Do some more projects ken
Working on it! I should have one coming out at the end of the month where I analyze my youtube comments. The projects take me a bit longer, and my real work has kind of been getting the better of me :(
@@KenJee_ds Thats great.. You got silver button and still you are replying us you are awesome man.
can you start a SLACK Channel for your community ?
I don't think I could manage multiple channels outside of the Discord. I recommend exploring that though! I will consider it if slack gets enough support!
@@KenJee_ds What's your Discord community ?
And what did you find out about your viewers? ;)
You will find out in about a month ;)
Ken is it Data or Data Hahah, you use both
jk
Thanks for the video
Maybe... it is both lol
2:47 hahah!
I hope it worked ;)
@@KenJee_ds It did!
Like in advance)
:)
How afraid should we be about choosing projects that end up failing? Sometimes I get caught in this idea that I need to find the perfect projects that I know would be successful before I even start exploring the data.
Don't be afraid! Even if the project outcome isn't exactly what you wanted, at least you learned something. Honestly, I don't think I've ever had a project go as I planned. What is important is that you adjust as you go and do your best to find something interesting! I hope this helps!
1:12-
1:44-
2:10-
2:50-
3:27-
4:10-
5:04-
5:52-
Can you become a data scientist without having a degree?
It would be extremely difficult, but I think it would be possible. To get in without a degree, you would have to do something extremely impressive on kaggle or in your own personal projects.
Any Special / Surprise Video for 100k SUB? :)))))