I was working on project which involves using deep learning for predicting cancer from images and CT scans (computer vision ) . Can you make a seperate series for Deep learning implmentation and understanding
@@farsbintanwar96 I don't do a ton of deep learning with my work, but I can make more videos on it for sure. I wouldn't say it is my strongest suit, but It would also probably help me to learn more if I made some videos on it.
Hi Ken! Thanks for the advices! I am currently working on a project that predicts used cars prices in my country using data from a popular car sales web site. Than comes a question: is there any problem to make a Linkedin or blog post of this work with this scraped data? All the data is data that you can visually check on the web site, but I am never sure about the use of data obtained by scraping. Thanks Ken :)
I'm scraping tweets and got to predict which one will become popular (the most number of retweets), I got access to the tweet content, number of following, followers of the writer etc... I'm struggling with it right now, if you know a project which seems similar I'd love to know about it to get some inspiration.
That last part of your video was top tier. For my capstone project in college, I had did a semester long data science project that had a ton of different parts. We collected our own data, cleaned it, and created new features out of existing ones. We didn’t have enough time to create a website or proper interface to interact with our results, but putting in that extra work is really what made the project stand out.
1.5 years ago I was just starting my sophomore year of undergrad, and I didn’t even understand half of what you said here lol. Now I’m at the end of my junior year and I feel like I could have a good conversation about this stuff. Feels good to grow!
@@KenJee_ds Yes, I think so too. I learn both data science and English from you, and you have a good pronunciation, easy for learning english (I'm russian). Thank you very much. You are the best v-logger about DS
Thank yoy for the very informative videos. I would add the following advice, which I wish someone had told me: while following courses or reading books in the Data Science learning process, we will be doing tiny projects. There will be a tendency to focus on the learning and not on the packaging of one's work. By packaging I mean: cleaning your code, properly commenting it, basically making it into a notebook you could publish, not into a code that just does the job. So, there is a tendency to produce drafts for these projects in the learning process, instead of final versions. And the assumption is that at that moment learing is the goal, and that at some later moment we will convert these into final portfolio projects, when we have more time that we can spare and dedicate to the cosmetics part. But then you end up a few of months later with a lot of small project drafts to finalize and the task of refactoring and documenting so many projects will seem daunting and you'll keep postponing it. Therefore, my advice is: dedicate some time during this learning process to finalizing and publishing your work, resist the temptation of pushing this task further into the future. There won't be a 'right' time.
Great video sir! For starters, it would be interesting for you to start posting videos on the specifics as you do the projects. That means, ‘learn by doing’. Most of the content out there is platitude and generalistic. Would be interesting if you started doing that.
Thanks for watching! I recommend taking a look at my data science project from scratch series: th-cam.com/play/PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t.html . In it I go through and do a whole project. I think this may meet your 'learn by doing criteria'. I hope this helps!
Great video dude. I think this is a topic that most learners are interested in and you approached it very well, all actionable information and no fluff.
Do not understand many of the terminologies used (I believe good insights to know what terms I will need to learn later) but at the same time, they made it a pretty informative video. Thank you for your guidance.
Hey man, Great page! I graduated with a degree in actuarial science. While I was fortunate to learn the rigorous of Linear Modeling, coding and finance in general, It did not prepare me with the skills to market my skills to employers.I think your doing great work here. Thank you!
11:39 "Flask Wrapper" - had to re-watch it a few times to pick up what he said there. Hope this helps others. (Flask is a Python web application framework).
What about data analysis specific projects? Initially, I thought I want to learn data science but then I've realized I enjoy the data analysis part more.
For these, I would leverage visualization tools like tableau and power BI. I would host the tableau projects on Tableau public. These are more about story telling, you should make a compelling point and illustrate it with different graphs. Just don't use pie charts haha
You spent almost half of your presentation talking about Machine learning(feature engineering, clustering, regression etc). In my humble opinion Data science has a lot to do with other aspects other than Machine learning. I realized many people focus a lot on machine learning which shrinks a lot of the aspects of data science. Kaggle, for example, encourages folks to hone their machine learning skills. Kaggle isn't a good place to learn full fledge data science skills. Data collection, Data Cleaning, Data manipulation, Exploratory Data analyses where all left out. Kaggle mostly brings ready-made datasets which limits many kagglers the ability to effectively develop their data science skills. Nevertheless I sincerely appreciate the value you have provided through your video. Thank you
Thanks for watching! I completely agree with you that data science is so much bigger than machine learning. Kaggle definitely has its shortcomings when it comes to truly learning the field. I tried to stress the importance of collecting your own data and telling a good story here. I agree that I probably didn't go into the other parts of the data science lifecycle as much as I should have. I would love to get your thoughts on my video: what does a data scientist actually do? I think I laid out the structure better there Thanks again for watching and the feedback! Best, Ken
So I just started my DS journey and I'm making my way through IBM's data science courses on digital nation. Since you've mentioned kaggle focusing more on ml, where would you reckon I refine my data understanding/preparation/exploration skills? Thanks
@@davidakinmade3523 Although kaggle can have its shortcomings in this area. I still think it is one of the best places to see this illustrated. For this I would look at datasets that are not for competition. There is some great exploratory analysis out there on datasets where the incentive is not to "optimize" for a solution.
I did a sentiment analysis where I compared several models, and logistic regression performed the best. I then scraped my own data from twitter for a predictive analysis.
I do a little bit of that in my kaggle project from scratch series! I would recommend checking that out. I will also try to make more content around that though!
Could you please make a video for master thesis in data science. As a masters student in my final semester I am struggling to find a topic. I got some idea how to go about but it would be great if you dedicate a video and also please explain some of the areas for research.
I recommend checking out this video: th-cam.com/video/yUrrf3Pm33s/w-d-xo.html&ab_channel=KenJee! I didn't have to do a thesis for my masters, so this video would likely be the best advice I could provide. I hope it helps!
Thanks a lot Ken! Can you please make a video on what software and programming skills are essential for a data scientist job? I am currently learning deeplearning in python. But I am still pretty weak in data visualization and plotting. What should I learn for data visualization? Which libraries in python or what other softwares should I learn to use?
Matplotlib and plotly are frequently used for vis! I also think it is reasonable to be comfortable with tableau or power BI. I will try to make a video on this!
@@KenJee_ds Thank you Ken! I am already using matplotlib, but I actually wanna know mastering what types of plots are the most essential? (i.e. histograms, bar charts etc)
Hi Ken, Thank you for your advice! This has been very helpful as an aspiring data scientist. I was wondering if I could get your thoughts on this project I am currently working on and if it is good enough to put on my portfolio. To quickly summarize, I did EDA on my Spotify listening history and merged it with a Spotify features dataset that listed features of each song (For example, tempo, major, acousticness, and valence are a couple examples). I wanted to see if there was a correlation between the songs I usually listen to and the features. However, there was very little correlation so I said that the model (using multiple linear regression) I would build would be very inaccurate at predicting, and showed the inaccuracy by building it and finding its mean squared error (which was very large; although there were some predictions that were close). In your opinion, would this suffice as a good project? If not, do you have any insight on how I could improve? Thank you for all your videos!
Hey Isac - Thanks for watching my video! That project sounds like a great one for your portfolio. The fact that you got your own spotify data and had an interesting reason for doing the project will go a long way. Projects don't always have to have "significant" findings to be useful for your resume. I would think of some theoretical "next steps" about how you could take this a step further to get build a better model. You don't necessarily have to go through with them (they may be too time or resource intensive), but you should know how you could improve. I hope this helps, and great work! Best, Ken
How would one categorize a project of something like "an algorithm to find an optimum point/value (either max or min) for a set of possible configurations" (for example a TSP-ish problem)...? Given that the problem requires that one understands the data from those configurations and the business value a solution would provide.
I work at a market research firm as an analytics client manager. I create many deliverables from our proprietary research/data, but I don’t know if I can share those publicly given NDAs. I’m not looking to leave at the moment, but for the future, do you have any thoughts on how to showcase my work? I guess I could maybe scrub the data, but I’m not sure how effective the story would come across at that point.
Scrubbing is an option. I think having them on your resume and being able to talk to them is sufficient though. If you are already in the domain and are applying , the project portfolio is less relevant
It’s so hard to do a project out of new idea 💡 feel like just quitting this field 10minths and still a noob in this field kills me from inside makes me wanna pull my hair Everytime I see a blank Jupiter notebook
I still feel like a noob quite often. It is part of this filed haha. I recommend trying to embrace the feeling, the struggle is what makes it fun! Maybe try the approach for project brainstorming that I highlight here: th-cam.com/video/MpF9HENQjDo/w-d-xo.html&ab_channel=KenJee
@@KenJee_ds Okay thanks, and thanks for making me find Kaggle, looks great and overwhelming at the same time lol It seems to me that ds and ml aren't one and another anymore, but have become so intertwined that you basically must learn both Or is it still possible to make a distinction between the two? are there jobs that require just one of them? Thank you
Hi Ken, loving the channel! I’m likely going to try and push into data science once my PhD is complete and I was wondering how big should a project be to make it worthy of going on a resume? I imagine something written in a few hours wouldn’t be worthy but where exactly this line falls I’m not so sure?
Thanks for watching and good luck finishing your Phd! What is it in? This may be surprising, but I wouldn't measure size by the amount of time or the lines of code it takes. I would measure it based on the size of the impact or how interesting the findings are. For example, I did an analysis of the Astros cheating scandal (premiering today), although it took me only a couple hours, it has a ton of relevancy. If it was on my resume, I am sure that people would ask about it. That probably isn't exactly what you were looking for haha. If you wan't a more black and white answer: If a project has all of the steps of the data science life-cycle, you should be fine putting it on your resume. I hope this helps -Ken
Hey Ken, great video ;) So I have done a project where ive scrapped data from a website, performed EDA and built a classifier analysing the performance of various ML model. How exactly do you suggest mentioning this on my resume? Like can you give me an example on how to go about writing it? Thanks
I would think of a good title for the project related to the topic. Next I would talk about the results of the project and then the methods you used. I have a video coming out Monday that goes into detail about this. It also provides a real example. Please stay tuned for that!
Hi Ken! First of all thank you for all the great videos that you are making. I just discovered your channel and I really appreciate it. As I have no internship or education directly link to programming I would like to create a Github page to show to the company that I'm still competent. I'm looking for a job as data analyst in the oil and gas area. Do you have some ideas about projects that I could do? Thank you in advance!
Hi Alex - Thanks for watching my videos! There are 2 main approaches for finding a good project: 1) Find a problem you are very interested in solving and go out and get the data for it 2) go on kaggle or google data sets and find some data that is interesting to you. In this video I go through my project brainstorming process: th-cam.com/video/MpF9HENQjDo/w-d-xo.html&ab_channel=KenJee . Unfortunately, I can't tell you what projects I would recommend because the best project is one you come up with based on your own interests.
I learned mostly through my own personal projects (on kaggle or other). I quickly found that the best way to improve model accuracy was to experiment with the features. In theory, there are infinite ways to engineer features, but only a finite amount of ways to tune a model.
@@KenJee_ds I see. Yeah feature engineering is kind of an open ended topic and often when I see clever ppl on kaggle, I think I'd never be able to think of that. What do you think is the fastest way to learn? Reading what others did and then try to recall it from memory?
@@Leon-pn6rb I think getting your hands dirty and experimenting is the best way to learn. There are a few out of the box techniques for feature engineering (pca, factor analysis, bagging, etc.) but you can really cut data however you want. If you have fun with it and explore constantly, you will begin to understand it intuitively. I recommend watching my data science project from scratch series, particularly the EDA portion of the analysis. You can see my thought process for picking it apart there.
Thanks very useful video, may i know for the person who is not from the IT background and if they want to start which software they should learn or in more demand, Salesforce, Tableau, QlikView or POWER-BI? I am very good in excelsheet and i like to find the correlation between two events or things. If you dont mind may i know from where should i start?
Thanks for watching! For starter projects, I recommend the ones I talk about in this video: th-cam.com/video/8igH8qZafpo/w-d-xo.html . For more advanced projects, you should find data that you are interested in. You can do this by browsing the kaggle datasets.
Hi, I feel like this topic is key for data science, I just wander why do you recommend Flask? could I use Django? or why not any other technology with python.
#32: Great great video, you are consistent with the message always!!! “Data Science is much about what’s going into the model as the actual selection and the things that you choose” wise words!!! #66daysofdata
Haha hopefully I don't sound too much like a broken record at this point 😂! Also, will be responding to your email about the prize tomorrow afternoon. I have to activate the free code for my course, so it will take 24 hrs!
I think any statistics book will do generally. This video has a few different free resources for learning the math: th-cam.com/video/zSwM5uVeylU/w-d-xo.html
Thanks for watching! I think you are referring to a "flask wrapper" that you would put around the model to make it an api endpoint. I believe that I was saying either doing that or scraping your own data are great ways to differentiate yourself. I hope this helps!
@@KenJee_ds yes sure. Thanks for quick response. I am UI developer. One more suggestion I would need . I see many TH-cam videos on same topics but they use different library and different way to finally achieve the things. How you suggest a beginner to tackle those things. As this for will sure confuse a new person .
I did some network analysis of Game of Thrones characters using Python to find out who is the most important character: www.linkedin.com/pulse/who-most-important-character-game-thrones-i-used-find-lukashevich-/ I'm very happy to watch Ken's content before starting my senior year of college. Not aiming for a DS title right away, looking for for BI Analytics.
Hi Tirupati - This video has 3 different projects that can be done th-cam.com/video/8igH8qZafpo/w-d-xo.html. In this one I talk about the projects that I did that helped me land a job: th-cam.com/video/imMPnCHvbkY/w-d-xo.html . I hope these help clarify things!
Any ideas on How I could begin choosing a topic for my Honours(Postgraduate) research (thesis) this year. It is in "Data Science for business decision support". I don't have any computer science background but I really would something interesting in this area because it is quite a broad area.
You could go through the data sets on kaggle and see what catches your eye, or you could go to business and ask if they have data that you could use / what data would be useful to them. I hope this helps!
A project can have data collection, data cleaning, exploratory data analysis, model building, and / or model deployment. I think if you do more than 2 of these in together it constitutes a project. I hope this helps!
Thanks for watching. Please watch my video The Best Free Data Science Courses Nobody is Talking about. Those are my recommendations for model building. th-cam.com/video/Ip50cXvpWY4/w-d-xo.html. I also have a playlist called data science fundamentals, where I do some very basic model building. I hope this helps!
find krish naik mate. he got it covered th-cam.com/video/bjsJOl8gz5k/w-d-xo.html Azure, flask , google cloud.. he would deploy in person step by step in the video. ---- from INDIA :)
I have Bank data contains customer Demographics including Account Type and Bank Balances. I want to know which customer to Target for Financing the bank products like Insurance, Autos etc
@@datascienceplayground2515 Excel would not be required, I generally recommend using python or R. You could do most of this analysis in excel if you wanted to though.
Hi, I work in a research company, can you please suggest me some projects related to Research or HR Domain as i m working in an HR Based Project Thanks
It is when you change the nature of the individual data points. For example you may have people's heights. You could convert that to a group of "tall people" and "short people". If you have geographic data points, you could convert them to distances from a common location.
Hello, I am from India and i am currently in class 12th. Now i have this very big doubt on weather i should take B.Tec in CSE or AI. I want to persue my career in AI and i will doing till PhD. What my parents thinks that i should take CSE for B.Tec, coz if i don't like AI then i would not be left with more options, and what my side is, that i just love AI, like i literally love it. So nowaday there are some colleges which offer a B.Tec course in AI. So i have compared the curriculums of Both AI and CSE and what i found is if yoh are clear that you want to do AI then definitely you should take AI in B.Tec. So i have this confusion of what to do in my B.Tec, can you please help me through this. It would be really a great help to me. My only consern is that if i take CSE then the topics like ML, AI, digital fabrication, AI and Humanity, DBMS, Computer networks, robotics, etc. I will miss this all as strong base, as direct in master i will be having less time, if i opt for AI then i have a preety good amount of time, and other side it's a wide spectrum open to me if i take CSE, so this is what my prospective is so what should i choose?
I think that either track you will do will be fine. Honestly you learn most of AI / ml through your own projects, so as long as you are doing that it really doesn't matter much what you learn in school. I promise you won't be behind in whichever you choose. If you want to do AI, I promise that you will learn plenty of generalizable skills as well so it isn't a real risk.
Thanks for watching! To me, advanced feature engineering is adding additional context to the data that you currently have. You can do this through ratios, grouping, or various packages. An example of this would be if you were analyzing a finance data set: You can calculate price / earnings ratio, sharpe ratio, or other metrics from the data that you already have. You can then build these into the model. A more basic example would be if you had basketball data and had the number of 3 points made and the number of 3 points taken. You can calculate the 3 point % make and build that into your model as well. A third example would be around text data, you can use the length of the text (of a tweet or something) as a feature, but you can also use sentiment, the number of verbs, etc. Hopefully these examples give you a bit more context!
@@KenJee_ds delighted by your super fast reply... Thanks a lot...I have just started doing data science for last 3-4 months and now looking for doing good projects...I found your content very good and I hope you keep helping us in future...thanks from INDIA
Hey Ken, AMAZING video!!! If I collect my own data, how can I reference it in the GitHub? Is it a good practice to add the CSV/txt file in the repository?
Thanks for watching! I am glad it was helpful to you! If the data file is manageable (less than 100 mb) I usually try to include it. If not, I would recommend posting the code that you used to scrape it (if you did). Hope this helps!
Thank you for ALL your recommendations. I'm studying Data analaysis under master course in South Korea. I have several questions below. 1. Do I need to follow the instructions in this video clip step by step? 2. Which is better to update portfolio between Github and Kaggle? 3. Which site could be a proper sameple to collect data at the beginning step?
Thank you for watching! 1) you can do the projects in any order that you like. I generally recommend doing linear regression and naive classifiers first then move on to the more complicated models. 2) I would recommend putting everything that you do on kaggle on your github. If you have other projects using data not from kaggle, you should put them all on your github. Generally, my git hub is more updated than my kaggle, but you should keep both recent if possible. 3) I personally like sports data, so an example would be a website like basketballreference.com you can also collect data through API's by using platforms like quantopian.com. I hope that this helps and answers your questions!
About 3-4 hours. There are some pretty good templates for wordpress websites out there. I had no experience with wordpress previously, and it was relatively straightforward to learn!
Hello everyone, Im a 4th year student and still dont have idea what project to do for my thesis. Anyone here have suggestions on timely projects i should do
@@KenJee_ds Oh my! Ken jee noticed me haha. Thaaanks ive been watching your videos for the past view days non stop, trying to prep myself for this coming schoolyear. Thanks for the tips.
Hi Logicc, thank you for watching the video! I think that discrete math, statistics / probability theory, linear algebra, and calculus are the relevant foundational math disciplines for data science.
Hi Ken, you recommended having some sort of data collection element involved in our projects but how would you show that? Other than including the datasets in your project files & explaining in a writeup, is there any way to show people what your data collection process is? Also, would you be able to make a video on working with big data / hadoop sometime in the future (if that's something you've had experience with)?
Thanks for watching Priyank! I recommend checking out my project from scratch series: th-cam.com/play/PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t.html . In this, I scrape data and show you how to document it. I've only worked with hadoop a little bit, so I don't think I will be doing a project on that too soon unfortunately though.
@@KenJee_ds Thanks Ken, I've been taking the kaggle micro courses you'd recommended to refresh my data science skills & that playlist is my next stop after finishing these courses! I recently found your channel and its evident that you put in a lot of effort & care into making quality content & connecting with your viewers (like answering all the comments). I really respect what you're doing & I'm glad to help support creators like you. Thanks!
This is great. I've always wanted to go into data science, but I keep hearing the whole "Go to college" deal, and honestly, I can't do it now (I did college and that was a failure).
Thanks for watching! It is a harder route without college, but if you can prove your value to a company through projects, I believe it is still possible. Best of luck!
Hi Ken, I am from india and I've just begun my journey to become a dats scientist. I'm not even really sure is data is the thing i love? I'm exploring my options and would do masters accordingly. Is it possible i can get in touch with you if I stumble upon something because you do look like you know what you're talking about. And also i subscribed 🔥
Thanks for watching and for the sub! Feel free to email me if you would like. My email address is in the about section of my channel. Good luck on your journey!
Thanks for watching everyone! I would love to hear about the projects you are working on in the comments section below!
I was working on project which involves using deep learning for predicting cancer from images and CT scans (computer vision ) . Can you make a seperate series for Deep learning implmentation and understanding
@@farsbintanwar96 I don't do a ton of deep learning with my work, but I can make more videos on it for sure. I wouldn't say it is my strongest suit, but It would also probably help me to learn more if I made some videos on it.
@@KenJee_ds of course definitely..after all we all are here to learn and grow together :)..all the best to us both in this case
Hi Ken! Thanks for the advices!
I am currently working on a project that predicts used cars prices in my country using data from a popular car sales web site.
Than comes a question: is there any problem to make a Linkedin or blog post of this work with this scraped data? All the data is data that you can visually check on the web site, but I am never sure about the use of data obtained by scraping.
Thanks Ken :)
I'm scraping tweets and got to predict which one will become popular (the most number of retweets), I got access to the tweet content, number of following, followers of the writer etc... I'm struggling with it right now, if you know a project which seems similar I'd love to know about it to get some inspiration.
That last part of your video was top tier. For my capstone project in college, I had did a semester long data science project that had a ton of different parts. We collected our own data, cleaned it, and created new features out of existing ones. We didn’t have enough time to create a website or proper interface to interact with our results, but putting in that extra work is really what made the project stand out.
Thanks for watching the video, and that's awesome! What was the project on?
I wish more schools would do projects like that!
@TOSAN I'm a beginner, I need to practice, can you share your projects, or GitHub that would help a lot.
Thank you.
1.5 years ago I was just starting my sophomore year of undergrad, and I didn’t even understand half of what you said here lol. Now I’m at the end of my junior year and I feel like I could have a good conversation about this stuff. Feels good to grow!
Hell yeah!
Wow thank you TH-cam algo! I’m an undergrad DS major and can’t wait to share with others in my major!
Thanks for watching. I hope my videos can be a useful resource to you! Please let me know if there are any other topics you would like me to cover!
just a forewarning data science is not usually an entry level job. those are analysts jobs, maybe a data enginner..
Dude you really helped spark a passion for data science in me. Thank you. The hardest part was always knowing where to start.
This is one of my favorite things to hear! Thanks for watching my videos!
wow! u actually took the time to answer every single question! what a boss
Doing my best haha!
I learn both data science and English from you. you have a fabulous pronunciation. Thank you🔥🔥🔥
Thank you! This is one of the best compliments I've ever gotten!
@@KenJee_ds Yes, I think so too. I learn both data science and English from you, and you have a good pronunciation, easy for learning english (I'm russian). Thank you very much. You are the best v-logger about DS
That data collection could really be the eye catching thing for recruiter. I am in. Thanks for the tips Ken.
Glad to hear the tips helped!
Thank yoy for the very informative videos.
I would add the following advice, which I wish someone had told me: while following courses or reading books in the Data Science learning process, we will be doing tiny projects. There will be a tendency to focus on the learning and not on the packaging of one's work. By packaging I mean: cleaning your code, properly commenting it, basically making it into a notebook you could publish, not into a code that just does the job.
So, there is a tendency to produce drafts for these projects in the learning process, instead of final versions. And the assumption is that at that moment learing is the goal, and that at some later moment we will convert these into final portfolio projects, when we have more time that we can spare and dedicate to the cosmetics part.
But then you end up a few of months later with a lot of small project drafts to finalize and the task of refactoring and documenting so many projects will seem daunting and you'll keep postponing it.
Therefore, my advice is: dedicate some time during this learning process to finalizing and publishing your work, resist the temptation of pushing this task further into the future. There won't be a 'right' time.
Great advice! Thank you for sharing!
Thank you for your videos. I'm starting at this, just taking notes of what you say and let's see how far I can get.
Thank you for watching them! I hope they have been helpful!
As a beginner, I'm extremely thankful for you. Your videos are super informative.
I'm thankful for you too! Thanks for watching my videos!
Great video sir! For starters, it would be interesting for you to start posting videos on the specifics as you do the projects. That means, ‘learn by doing’. Most of the content out there is platitude and generalistic. Would be interesting if you started doing that.
Thanks for watching! I recommend taking a look at my data science project from scratch series: th-cam.com/play/PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t.html . In it I go through and do a whole project. I think this may meet your 'learn by doing criteria'. I hope this helps!
Great video dude. I think this is a topic that most learners are interested in and you approached it very well, all actionable information and no fluff.
Thanks for watching! I appreciate the kind feedback!
I’m starting on projects to build my portfolio. Happy that I watched this video on time Thanks Ken😇
Awesome! Thanks for watching!
Do not understand many of the terminologies used (I believe good insights to know what terms I will need to learn later) but at the same time, they made it a pretty informative video. Thank you for your guidance.
Thanks for watching! I promise the terms will come as you learn more!
@@KenJee_ds Thank you for sharing this!
Hey man, Great page!
I graduated with a degree in actuarial science.
While I was fortunate to learn the rigorous of Linear Modeling, coding and finance in general, It did not prepare me with the skills to market my skills to employers.I think your doing great work here. Thank you!
Very useful video Ken, thanks for taking the time to create it
Thanks for taking the time to watch it Elias!!
Nice detailing, Ken!
Thanks!
This is a great video, plenty of insight with direction. Thank you.
Glad you enjoyed it Roy! Thanks for watching!
Thanks for the super useful video. You might want to check the audio as the audio level varied a lot during the video
Thanks for watching and for the feedback! Hopefully have improved the audio in my more recent videos!
Great advice bro, had to give it a like and subscribed
Thanks for watching and for the sub! Glad you liked it!
It's pretty helpful for me, thank you very much.
Glad to hear! Thanks for watching!
11:39 "Flask Wrapper" - had to re-watch it a few times to pick up what he said there. Hope this helps others. (Flask is a Python web application framework).
Yes! Thank you for the clarification, I wasn't very clear here
good stuff. prowling your page for some walk throughs too
Thanks for watching! Hopefully they will be useful to you!
Thanks for this video.. It was really helpful. 👌
Thanks for watching!
This is really tons of value
Glad to hear! Thanks for watching Javier!
Super useful! Thanks!
Thanks for watching!
thank you. it was really helpful
Glad I could help!
What about data analysis specific projects? Initially, I thought I want to learn data science but then I've realized I enjoy the data analysis part more.
For these, I would leverage visualization tools like tableau and power BI. I would host the tableau projects on Tableau public. These are more about story telling, you should make a compelling point and illustrate it with different graphs. Just don't use pie charts haha
@@KenJee_ds Cheers!
Thanks, great advice!
No problem!
You spent almost half of your presentation talking about Machine learning(feature engineering, clustering, regression etc). In my humble opinion Data science has a lot to do with other aspects other than Machine learning. I realized many people focus a lot on machine learning which shrinks a lot of the aspects of data science. Kaggle, for example, encourages folks to hone their machine learning skills. Kaggle isn't a good place to learn full fledge data science skills. Data collection, Data Cleaning, Data manipulation, Exploratory Data analyses where all left out. Kaggle mostly brings ready-made datasets which limits many kagglers the ability to effectively develop their data science skills. Nevertheless I sincerely appreciate the value you have provided through your video. Thank you
Thanks for watching! I completely agree with you that data science is so much bigger than machine learning. Kaggle definitely has its shortcomings when it comes to truly learning the field. I tried to stress the importance of collecting your own data and telling a good story here. I agree that I probably didn't go into the other parts of the data science lifecycle as much as I should have. I would love to get your thoughts on my video: what does a data scientist actually do? I think I laid out the structure better there
Thanks again for watching and the feedback!
Best,
Ken
@@KenJee_ds Thank you for your reply. I will check in with your other video.
So I just started my DS journey and I'm making my way through IBM's data science courses on digital nation. Since you've mentioned kaggle focusing more on ml, where would you reckon I refine my data understanding/preparation/exploration skills? Thanks
@@davidakinmade3523 Although kaggle can have its shortcomings in this area. I still think it is one of the best places to see this illustrated. For this I would look at datasets that are not for competition. There is some great exploratory analysis out there on datasets where the incentive is not to "optimize" for a solution.
Hey Ken,
Could you elaborate more on last 3 points in your next video!
I can definitely do that!
Thanks Ken.
Thanks for watching!
I did a sentiment analysis where I compared several models, and logistic regression performed the best. I then scraped my own data from twitter for a predictive analysis.
Awesome stuff!
@@KenJee_ds Thanks, currently applying to internships and MSc programmes.
Thank you so much
This is really helpful
Glad it was helpful! Thanks for watching!
You doing excellent job Sir.
Thanks for watching Sean!
Thanks for the information. Can you please make a video on feature engineering. How to convert raw data to meaningful data.
I do a little bit of that in my kaggle project from scratch series! I would recommend checking that out. I will also try to make more content around that though!
Could you please make a video for master thesis in data science. As a masters student in my final semester I am struggling to find a topic. I got some idea how to go about but it would be great if you dedicate a video and also please explain some of the areas for research.
I recommend checking out this video: th-cam.com/video/yUrrf3Pm33s/w-d-xo.html&ab_channel=KenJee! I didn't have to do a thesis for my masters, so this video would likely be the best advice I could provide. I hope it helps!
@@KenJee_ds Thank you!
Thanks a lot Ken! Can you please make a video on what software and programming skills are essential for a data scientist job? I am currently learning deeplearning in python. But I am still pretty weak in data visualization and plotting. What should I learn for data visualization? Which libraries in python or what other softwares should I learn to use?
Matplotlib and plotly are frequently used for vis! I also think it is reasonable to be comfortable with tableau or power BI. I will try to make a video on this!
@@KenJee_ds Thank you Ken! I am already using matplotlib, but I actually wanna know mastering what types of plots are the most essential? (i.e. histograms, bar charts etc)
@@shahrinnakkhatra2857 Ohhh. Let me think on a resource for that!
Hi Ken,
Thank you for your advice! This has been very helpful as an aspiring data scientist. I was wondering if I could get your thoughts on this project I am currently working on and if it is good enough to put on my portfolio. To quickly summarize, I did EDA on my Spotify listening history and merged it with a Spotify features dataset that listed features of each song (For example, tempo, major, acousticness, and valence are a couple examples). I wanted to see if there was a correlation between the songs I usually listen to and the features. However, there was very little correlation so I said that the model (using multiple linear regression) I would build would be very inaccurate at predicting, and showed the inaccuracy by building it and finding its mean squared error (which was very large; although there were some predictions that were close).
In your opinion, would this suffice as a good project? If not, do you have any insight on how I could improve?
Thank you for all your videos!
Hey Isac - Thanks for watching my video! That project sounds like a great one for your portfolio. The fact that you got your own spotify data and had an interesting reason for doing the project will go a long way. Projects don't always have to have "significant" findings to be useful for your resume. I would think of some theoretical "next steps" about how you could take this a step further to get build a better model. You don't necessarily have to go through with them (they may be too time or resource intensive), but you should know how you could improve. I hope this helps, and great work!
Best,
Ken
@@KenJee_ds Oo that's a great idea THANK YOU! I really appreciate you and your videos.
@@isaclee9919 Happy to help and good luck with the project!
you look like a boss!!!
11:49 wow I literally just learned how to do this for one of my old projects.
Awesome!
How would one categorize a project of something like "an algorithm to find an optimum point/value (either max or min) for a set of possible configurations" (for example a TSP-ish problem)...? Given that the problem requires that one understands the data from those configurations and the business value a solution would provide.
I work at a market research firm as an analytics client manager. I create many deliverables from our proprietary research/data, but I don’t know if I can share those publicly given NDAs. I’m not looking to leave at the moment, but for the future, do you have any thoughts on how to showcase my work? I guess I could maybe scrub the data, but I’m not sure how effective the story would come across at that point.
Scrubbing is an option. I think having them on your resume and being able to talk to them is sufficient though. If you are already in the domain and are applying , the project portfolio is less relevant
it is difficult to concentrate with the background music, although your content is good :)
Thanks for the feedback. I have reduced the background music in more recent videos!
It’s so hard to do a project out of new idea 💡 feel like just quitting this field 10minths and still a noob in this field kills me from inside makes me wanna pull my hair Everytime I see a blank Jupiter notebook
I still feel like a noob quite often. It is part of this filed haha. I recommend trying to embrace the feeling, the struggle is what makes it fun! Maybe try the approach for project brainstorming that I highlight here: th-cam.com/video/MpF9HENQjDo/w-d-xo.html&ab_channel=KenJee
Great video, thank you, but these sound more like strictly machine learning than data science tbh
Thanks for watching! That is fair, I do think each project should have a large EDA phase, I think that is more of where the ds comes in.
@@KenJee_ds Okay thanks, and thanks for making me find Kaggle, looks great and overwhelming at the same time lol
It seems to me that ds and ml aren't one and another anymore, but have become so intertwined that you basically must learn both
Or is it still possible to make a distinction between the two? are there jobs that require just one of them?
Thank you
Ken please make tutorials video on specific Data Science projects....it would be helpful
It will take me some time, but I will do an end-to-end data science project sometime in the next 3-4 months! Thanks for watching!
@@KenJee_ds thanks.......actually i need it for final year.....
Hi Ken, loving the channel! I’m likely going to try and push into data science once my PhD is complete and I was wondering how big should a project be to make it worthy of going on a resume? I imagine something written in a few hours wouldn’t be worthy but where exactly this line falls I’m not so sure?
Thanks for watching and good luck finishing your Phd! What is it in? This may be surprising, but I wouldn't measure size by the amount of time or the lines of code it takes. I would measure it based on the size of the impact or how interesting the findings are. For example, I did an analysis of the Astros cheating scandal (premiering today), although it took me only a couple hours, it has a ton of relevancy. If it was on my resume, I am sure that people would ask about it.
That probably isn't exactly what you were looking for haha. If you wan't a more black and white answer: If a project has all of the steps of the data science life-cycle, you should be fine putting it on your resume.
I hope this helps -Ken
What projects do you recommend Data Analysts to do? & is there a website I can use to build these projects?
kaggle is amazing for projects and you can see how other people did them if youre stuck
As Alex said, kaggle is great. I also recommend checking out Tableau's makeover Monday. This is a great opportunity for flexing visualization skills!
Hey Ken, great video ;)
So I have done a project where ive scrapped data from a website, performed EDA and built a classifier analysing the performance of various ML model. How exactly do you suggest mentioning this on my resume? Like can you give me an example on how to go about writing it?
Thanks
I would think of a good title for the project related to the topic. Next I would talk about the results of the project and then the methods you used. I have a video coming out Monday that goes into detail about this. It also provides a real example. Please stay tuned for that!
@@KenJee_ds Sure, cheers ;)
Hi Ken! First of all thank you for all the great videos that you are making. I just discovered your channel and I really appreciate it.
As I have no internship or education directly link to programming I would like to create a Github page to show to the company that I'm still competent. I'm looking for a job as data analyst in the oil and gas area. Do you have some ideas about projects that I could do? Thank you in advance!
Hi Alex - Thanks for watching my videos! There are 2 main approaches for finding a good project: 1) Find a problem you are very interested in solving and go out and get the data for it 2) go on kaggle or google data sets and find some data that is interesting to you. In this video I go through my project brainstorming process: th-cam.com/video/MpF9HENQjDo/w-d-xo.html&ab_channel=KenJee . Unfortunately, I can't tell you what projects I would recommend because the best project is one you come up with based on your own interests.
How did you learn complicated feature engineering? Through reading kaggle kernels?
I learned mostly through my own personal projects (on kaggle or other). I quickly found that the best way to improve model accuracy was to experiment with the features. In theory, there are infinite ways to engineer features, but only a finite amount of ways to tune a model.
@@KenJee_ds I see. Yeah feature engineering is kind of an open ended topic and often when I see clever ppl on kaggle, I think I'd never be able to think of that.
What do you think is the fastest way to learn? Reading what others did and then try to recall it from memory?
@@Leon-pn6rb I think getting your hands dirty and experimenting is the best way to learn. There are a few out of the box techniques for feature engineering (pca, factor analysis, bagging, etc.) but you can really cut data however you want. If you have fun with it and explore constantly, you will begin to understand it intuitively. I recommend watching my data science project from scratch series, particularly the EDA portion of the analysis. You can see my thought process for picking it apart there.
@@KenJee_ds Will do! Thank you for being so helpful. I really need some guidance in this crucial time for my career. Really appreciate it.
Thanks very useful video, may i know for the person who is not from the IT background and if they want to start which software they should learn or in more demand, Salesforce, Tableau, QlikView or POWER-BI?
I am very good in excelsheet and i like to find the correlation between two events or things. If you dont mind may i know from where should i start?
I think tableau and power bi are good ones to start with. I would also explore learning some python
@@KenJee_ds thank you very much for your response
Thanks Ken for these amazing videos.
Could you please tell us datasets for these projects from Kaggle.
Thanks for watching! For starter projects, I recommend the ones I talk about in this video: th-cam.com/video/8igH8qZafpo/w-d-xo.html . For more advanced projects, you should find data that you are interested in. You can do this by browsing the kaggle datasets.
Hi, I feel like this topic is key for data science, I just wander why do you recommend Flask? could I use Django? or why not any other technology with python.
Flask is just the first that came to mind and it was the easiest way to do this for me. Now I generally recommend streamlit!
Can you recommend any of the datasets in kaggle to begin with.
I recommend the ones I mention in this video: th-cam.com/video/8igH8qZafpo/w-d-xo.html&ab_channel=KenJee
Can you recommend any coding books that aid in doing these types of supervised and unsupervised models?
I personally like this book quite a bit! amzn.to/39y1fkM
#32: Great great video, you are consistent with the message always!!!
“Data Science is much about what’s going into the model as the actual selection and the things that you choose” wise words!!! #66daysofdata
Haha hopefully I don't sound too much like a broken record at this point 😂! Also, will be responding to your email about the prize tomorrow afternoon. I have to activate the free code for my course, so it will take 24 hrs!
Which statistics book would you recommend for data Science
I think any statistics book will do generally. This video has a few different free resources for learning the math: th-cam.com/video/zSwM5uVeylU/w-d-xo.html
hi, future me, let me see where you at now! (wish you can find this comment when you back here)
a recommendation to improve videos, please increase your voice or get a stronger mic. thank you
Done, thank you for the feedback!
Sounds like clustering models (unsupervised) can be used to design database tables
And help with normalization
Haven't seen this use case, but I guess it could be!
What's the "fla..scraper" thing you mentioned to automate the scraping? Loved the video
Thanks for watching! I think you are referring to a "flask wrapper" that you would put around the model to make it an api endpoint. I believe that I was saying either doing that or scraping your own data are great ways to differentiate yourself. I hope this helps!
@@KenJee_ds Oh alright, thanks!
@ Ken Jee is it okay to dive directly into coding and then googling one by one to the understand the usage and the math behind the algorithm.
I think that is totally fine! Make sure to go back and learn the math at some point though!
@@KenJee_ds yes sure.
Thanks for quick response.
I am UI developer.
One more suggestion I would need . I see many TH-cam videos on same topics but they use different library and different way to finally achieve the things. How you suggest a beginner to tackle those things. As this for will sure confuse a new person .
I did some network analysis of Game of Thrones characters using Python to find out who is the most important character:
www.linkedin.com/pulse/who-most-important-character-game-thrones-i-used-find-lukashevich-/
I'm very happy to watch Ken's content before starting my senior year of college.
Not aiming for a DS title right away, looking for for BI Analytics.
I really enjoyed the analysis!
Hi, thanks for your video. I wish there were English subtitles...for non native like me.
I will try to work on including them! Thank you for watching!
Please explain any of the machine learning project
Hi Tirupati - This video has 3 different projects that can be done th-cam.com/video/8igH8qZafpo/w-d-xo.html. In this one I talk about the projects that I did that helped me land a job: th-cam.com/video/imMPnCHvbkY/w-d-xo.html . I hope these help clarify things!
@@KenJee_ds I just got past the python basics and I feel like I'm so far behind in my learning.
Any ideas on How I could begin choosing a topic for my Honours(Postgraduate) research (thesis) this year. It is in "Data Science for business decision support".
I don't have any computer science background but I really would something interesting in this area because it is quite a broad area.
You could go through the data sets on kaggle and see what catches your eye, or you could go to business and ask if they have data that you could use / what data would be useful to them. I hope this helps!
Good video
Thanks for watching Shubham!
what is considered as a project in data science industry ??
A project can have data collection, data cleaning, exploratory data analysis, model building, and / or model deployment. I think if you do more than 2 of these in together it constitutes a project. I hope this helps!
@@KenJee_ds thank you Ken.. You are awesome... Keep coming with ds stuff 💙
Can you please suggest some tutorials for learning deployment of Machine learning models
Thanks for watching. Please watch my video The Best Free Data Science Courses Nobody is Talking about. Those are my recommendations for model building. th-cam.com/video/Ip50cXvpWY4/w-d-xo.html. I also have a playlist called data science fundamentals, where I do some very basic model building. I hope this helps!
find krish naik mate. he got it covered th-cam.com/video/bjsJOl8gz5k/w-d-xo.html
Azure, flask , google cloud.. he would deploy in person step by step in the video.
---- from INDIA :)
I have Bank data contains customer Demographics including Account Type and Bank Balances. I want to know which customer to Target for Financing the bank products like Insurance, Autos etc
That sounds like a good data set! It sounds like a classification or clustering project to me!
Do I require to Excel for Modelling this data?
@@datascienceplayground2515 Excel would not be required, I generally recommend using python or R. You could do most of this analysis in excel if you wanted to though.
Hi,
I work in a research company, can you please suggest me some projects related to Research or HR Domain as i m working in an HR Based Project
Thanks
Great video..... I want to do a project related to financial institution could you please recommend any?
The most common one is to predict stock or bitcoin prices using a recurrent neural net. I would recommend exploring that
@@KenJee_ds thanks jee....
What feature engineering actually means?
It is when you change the nature of the individual data points. For example you may have people's heights. You could convert that to a group of "tall people"
and "short people". If you have geographic data points, you could convert them to distances from a common location.
@@KenJee_ds Awesome.. Thanks man :)
Hello,
I am from India and i am currently in class 12th. Now i have this very big doubt on weather i should take B.Tec in CSE or AI. I want to persue my career in AI and i will doing till PhD. What my parents thinks that i should take CSE for B.Tec, coz if i don't like AI then i would not be left with more options, and what my side is, that i just love AI, like i literally love it. So nowaday there are some colleges which offer a B.Tec course in AI. So i have compared the curriculums of Both AI and CSE and what i found is if yoh are clear that you want to do AI then definitely you should take AI in B.Tec. So i have this confusion of what to do in my B.Tec, can you please help me through this. It would be really a great help to me.
My only consern is that if i take CSE then the topics like ML, AI, digital fabrication, AI and Humanity, DBMS, Computer networks, robotics, etc. I will miss this all as strong base, as direct in master i will be having less time, if i opt for AI then i have a preety good amount of time, and other side it's a wide spectrum open to me if i take CSE, so this is what my prospective is so what should i choose?
I think that either track you will do will be fine. Honestly you learn most of AI / ml through your own projects, so as long as you are doing that it really doesn't matter much what you learn in school. I promise you won't be behind in whichever you choose. If you want to do AI, I promise that you will learn plenty of generalizable skills as well so it isn't a real risk.
@@KenJee_ds Thank you very much 😄😄😄
Can you please describe broadly about advanced feature engineering?
Thanks for watching! To me, advanced feature engineering is adding additional context to the data that you currently have. You can do this through ratios, grouping, or various packages. An example of this would be if you were analyzing a finance data set: You can calculate price / earnings ratio, sharpe ratio, or other metrics from the data that you already have. You can then build these into the model. A more basic example would be if you had basketball data and had the number of 3 points made and the number of 3 points taken. You can calculate the 3 point % make and build that into your model as well. A third example would be around text data, you can use the length of the text (of a tweet or something) as a feature, but you can also use sentiment, the number of verbs, etc.
Hopefully these examples give you a bit more context!
@@KenJee_ds delighted by your super fast reply... Thanks a lot...I have just started doing data science for last 3-4 months and now looking for doing good projects...I found your content very good and I hope you keep helping us in future...thanks from INDIA
@@anindyasadhukhan7139 Happy to help! I am glad the content has been useful to you!
Hey Ken,
AMAZING video!!!
If I collect my own data, how can I reference it in the GitHub? Is it a good practice to add the CSV/txt file in the repository?
Thanks for watching! I am glad it was helpful to you! If the data file is manageable (less than 100 mb) I usually try to include it. If not, I would recommend posting the code that you used to scrape it (if you did).
Hope this helps!
@@KenJee_ds very clear, thanks
Thank you for ALL your recommendations. I'm studying Data analaysis under master course in South Korea. I have several questions below.
1. Do I need to follow the instructions in this video clip step by step?
2. Which is better to update portfolio between Github and Kaggle?
3. Which site could be a proper sameple to collect data at the beginning step?
Thank you for watching!
1) you can do the projects in any order that you like. I generally recommend doing linear regression and naive classifiers first then move on to the more complicated models.
2) I would recommend putting everything that you do on kaggle on your github. If you have other projects using data not from kaggle, you should put them all on your github. Generally, my git hub is more updated than my kaggle, but you should keep both recent if possible.
3) I personally like sports data, so an example would be a website like basketballreference.com you can also collect data through API's by using platforms like quantopian.com.
I hope that this helps and answers your questions!
Wow!! I really appreciated suprisingly quick response!
I hope you to upload a various informative video clip related to Data Analysis :)
@@wsdclub I have some coming in the next few weeks!
how long did it take u to make ur own website
About 3-4 hours. There are some pretty good templates for wordpress websites out there. I had no experience with wordpress previously, and it was relatively straightforward to learn!
Hello everyone, Im a 4th year student and still dont have idea what project to do for my thesis. Anyone here have suggestions on timely projects i should do
Im having a hard time thinking a "unique" project
I think this video may help! th-cam.com/video/yUrrf3Pm33s/w-d-xo.html&ab_channel=KenJee
@@KenJee_ds Oh my! Ken jee noticed me haha. Thaaanks ive been watching your videos for the past view days non stop, trying to prep myself for this coming schoolyear. Thanks for the tips.
Love this shit
Glad to hear! Thanks for watching!
hey Ken can you plzz suggest me some data science projects
I'll try.
Hey Ken, what kind of math requirements are needed for data scientist?
Hi Logicc, thank you for watching the video! I think that discrete math, statistics / probability theory, linear algebra, and calculus are the relevant foundational math disciplines for data science.
Comment first, then watch!
I am an undergraduate math major student. Could you please recommend some project ideas related to mathematics?
I recommend going on kaggle.com and finding a dataset that interests you. I would argue that all data science projects are related to math haha
Bro work on the sound level it's too low
Plz increase your sound volumes
Will do!
Hi Ken, you recommended having some sort of data collection element involved in our projects but how would you show that? Other than including the datasets in your project files & explaining in a writeup, is there any way to show people what your data collection process is?
Also, would you be able to make a video on working with big data / hadoop sometime in the future (if that's something you've had experience with)?
Thanks for watching Priyank! I recommend checking out my project from scratch series: th-cam.com/play/PL2zq7klxX5ASFejJj80ob9ZAnBHdz5O1t.html . In this, I scrape data and show you how to document it.
I've only worked with hadoop a little bit, so I don't think I will be doing a project on that too soon unfortunately though.
@@KenJee_ds Thanks Ken, I've been taking the kaggle micro courses you'd recommended to refresh my data science skills & that playlist is my next stop after finishing these courses!
I recently found your channel and its evident that you put in a lot of effort & care into making quality content & connecting with your viewers (like answering all the comments).
I really respect what you're doing & I'm glad to help support creators like you. Thanks!
This is great. I've always wanted to go into data science, but I keep hearing the whole "Go to college" deal, and honestly, I can't do it now (I did college and that was a failure).
Thanks for watching! It is a harder route without college, but if you can prove your value to a company through projects, I believe it is still possible. Best of luck!
Hi Ken, I am from india and I've just begun my journey to become a dats scientist. I'm not even really sure is data is the thing i love? I'm exploring my options and would do masters accordingly. Is it possible i can get in touch with you if I stumble upon something because you do look like you know what you're talking about. And also i subscribed 🔥
Thanks for watching and for the sub! Feel free to email me if you would like. My email address is in the about section of my channel. Good luck on your journey!
Don’t forget the ice cream! Lol
Are you a data scientist
I am a director of data science, but I still do hands on data science work about 50% of the time.
How can I become a data scientist
@@thelilpippin do hiphop
Fight!
1:16 Thank me later
thanks
🔥🔥
She s daughter of the governor
so long videos...
You said Everything 🤣
haha