Roasting My First Data Science Project (Common Project Mistakes)

[@KenJee_ds]

Thanks for watching everyone! I would love to hear what projects you're working on in the comments section below! If you want some more project advice, check out my project beginners playlist! kzbin.info/aero/PL2zq7klxX5ASt4dLSAd2FMoY3Og3V0jZv

[@DataProfessor]

Glad to see you Ken, thanks for sharing your learning experience of your first project. This video is so good, gotta take some notes: 1. Start that first project, rinse and repeat 2. Wrestle with your problem and don’t give up too early 3. Don’t be afraid to make mistakes, experiment and learn from it 4. Embrace failure and take it as a learning opportunity 5. Ken’s tips: project planning, problem definition, documentation, clear communication and sharing your work

[@anubhavgoyal3211]

Thanku for the notes:)

[@DataProfessor]

@@anubhavgoyal3211 My pleasure! 😆

[@Ibraheem_ElAnsari]

Nice to see you Prof 😀 !

[@KenJee_ds]

Thanks for watching data prof! If I had your channel back then I wouldn't have made any of these mistakes 😆

[@DataProfessor]

@@KenJee_ds Wow, that’s very kind of you, I can say the same for your channel, packed with tips, tricks, entertainment and papayas for all data enthusiasts 😊

[@jaemms2957]

Thanks for this Ken. The key takeaway, " it's important to wrestle with your data and not give up"

[@KenJee_ds]

True!!

[@alittleofthis08]

Thank you so much for your great content, Ken! I enjoy your channel so much. I'm kind of an outlier here. LOL I did a masters in statistics in the 80's and then was a statistical consultant for agricultural projects using SAS for several years. After a 35-year hiatus from formal data analysis (but not from thinking data and teaching math) I came back last year to complete a data science certification to try to get up to speed in the data world. Wow! What a challenge! I did one big project for my certification, but I want to do several more for my own benefit and for my portfolio before I start looking for contracts. Your videos give me lots of possibilities. I just really enjoy how you present ideas. Thank you again.

[@KenJee_ds]

Awesome! Looking forward to hearing more about your future projects!

[@datasciencefromzero2666]

Failures and bad experiences are what I need to hear as a junior. Because that's what happens to me mostly and everyone talks about good things happened to them which is unreal.

[@Mhurilo10]

I'm not a data analyst or anything but just the idea of having an automated judge was enough to make me watch the whole thing haha

[@KenJee_ds]

Thanks for watching!!

[@harshitsati]

I have really lost count of all the data science mini projects I've done, now I need to do a major one :life:

[@KenJee_ds]

hahahah

[@piyushzope10]

Thanks for sharing your experience... Data science is all about keep going with lots variation until we get results which can make impact on our purpose.

[@KenJee_ds]

Thanks for watching Piyush!

[@NicholasRenotte]

HE'S BACK!! Awesome vid Ken!

[@NicholasRenotte]

Also, how can I get on the list @ 3:25?

[@arnavaggarwal1940]

He is back 🔥🔥

[@KenJee_ds]

Back again!

[@ShubhamGhuleCodes]

Such useful insights for beginners! Thanks, Ken🙌

[@KenJee_ds]

Thanks for watching my video!!

[@Ibraheem_ElAnsari]

Gosh even though we have the podcast this type of videos hits differently. Glad to see u back here Ken ! Also 2:08 was hilarious haha .. I would love to work on a cooking project for my next one, I'll share it with the community when I'm done :)

[@piyushzope10]

I'm just curious...Cooking project? What is that? 🤔

[@KenJee_ds]

Glad to be back making this type of video again!

[@thinamG]

Thank you so much for sharing, Ken.

[@KenJee_ds]

Thanks for watching!!

[@jimiknewstubb]

Keep ‘em coming Ken! Cheers from NZ

[@KenJee_ds]

Will do my best!!

[@mouniaelkoraichi2502]

Just what I needed to hear 😊

[@KenJee_ds]

Excellent!

[@piyushzope10]

Hey welcome back... I must say... The quality of video improved alot... The editing and all other things are seems more better 🙌

[@KenJee_ds]

Thanks Piyush! I hired an editor, and he is killing it!

[@piyushzope10]

@@KenJee_ds Ohhhh... That's good. Now you can spend more time on focusing on the content...

[@peeintea]

That's exactly the thing i am facing right now, thanks.

[@KenJee_ds]

I hope it helps!

[@mattRRgraham1996]

Lol, for my research project in grad school I wrote some models based off of LSA, GloVe, and Word2Vec and the best accuracy I got was 16% from w2v. My LSA model was scoring at 4% lmao. it was hilariously bad. It really was humorous to me just how bad they were performing. I was able to (over some grueling months) get my W2V model up to 65% accuracy which for the task at hand, I still consider to be quite good honestly.

[@KenJee_ds]

In theory could you just do the opposite of what your models were saying and do really well 😂

[@datasciencegraduate432]

Great video Ken! Can definitely relate to your thumbnail when I was working on my first project lol 😂

[@anubhavgoyal3211]

Welcome Back! It's been a long time😀

[@KenJee_ds]

💪

[@DakotahTWilson]

Thanks Ken! I'm just starting my first data project for grad school and it's still in the data collection phase. I have a team of four undergrads working for me so right now I'm still trying to figure out what it is I'll want to analyze and present to the grant makers but we have some time to figure that part out. Do you have any advice for the pre-planning phase of research? I feel more comfortable in my coding than I do in project management but in this case, it's all part of the same project.

[@KenJee_ds]

I would spend a lot of time asking questions of the different topic areas that you could have. I usually ask, who would this analysis impact? What changes could be made as a result of this? Is the data available? Is the data usable? I think good pm just comes down to asking good questions

[@MrViperHiggins]

Great video Ken, it's something I struggle with too having a specific goal in mind. Its the most important thing in any project. Also love the Ryder Cup shirt Europe will come back swinging next time lol!

[@829Alejandro]

Hey Ken, I've been enjoying your videos! Great channel you have by the way. I'm currently a junior in college, pursuing a degree in mathematics and I have a interest in the data science field. I've taken a couple of programming classes like intro to computer science and predicative analytics and I still plan to learn more on my own free time. Will employers still prefer a computer science major over a math major? What are your thoughts?

[@farhanislam8463]

Is most code for almost all projects just copied from one source or another, and we just change the parameter values ,like epochs, or batch_size? Or is there something else, regarding making projects?

[@KenJee_ds]

A good portion of it is in the model building phase. I think some of the exploration code should vary quite a bit depending on the data you're using

[@farhanislam8463]

@@KenJee_ds Thanks. Got it.

[@a.hsumon9640]

Hi Ken. I am a undergrad student and I am deeply interested in Data science. In final year, should I do thesis or do project on data science? ThanksInAdvance

[@KenJee_ds]

I personally would recommend a project! I don't think a thesis would set you back or anything though

[@a.hsumon9640]

@@KenJee_ds Thanks, man. I appreciate your work. Keep it up.

[@SeattleDataGuy]

I love the MMA data set. Also, MMA math never works.

[@KenJee_ds]

Lol so true!

[@SeattleDataGuy]

@@KenJee_ds Sounds like their needs to be a data nerds ufc watch part some day in the future

[@arthurmiller5276]

🤔"If you knew then what you know now" (or even recreate the same project now) could you do a better job or is predicting the judging on a match just a 50% success? Also, did your model take into consideration the judges themselves as (in)consistent variables? Everyone's first attempts at something don't look good as subsequent efforts, particularly after years of experience. Do you apply how you've "failed" (even today) to improving your skills or possibly utilize the failure data as part of your model(s)?

[@federiconicolasgomez1928]

Yeeeaaa here again¡¡¡

[@KenJee_ds]

Yes!!

[@TinaHuang1]

ah I thought you were gonna say you never did it :x

[@KenJee_ds]

Lol I finished! Have more faith in me!

[@shrutijain1628]

But talkin ✨g about something purely my idea . I think I had problem with defining the problem statement and looking for right data . I just use to see the topic I want to work on and not on the problem statement. And secondly also I use to just take data if it's related to the topic and not to the problem statement biggest mistake . Took me a lot of time to get out of this mindset

[@KenJee_ds]

Glad you're out of it now Shruti!