Video

The p-value, is the probability that the observed data occur if the null hypothesis is true

No, I don't. But it is a great idea for a video. Thank you

@@karinadatascientist thanks karina ♥️

yes, my email is karinadatascientist@gmail.com

@@karinadatascientist thanks.. Will send you mails.

I look at that as an additional tool that will be must use in most professions rather than replacement. It can definitely speed up your work, but it can't replace your brain, ideas, vision and logic

Thanks

Great to hear!

Wow, you follow me in 2 platforms, that's cool. Thank you! There are jobs for "Sustainability data analysts" in electricity companies, airlines and all the renewables (solar, wind energy). I actually think it is a great field to be in, even maybe too early, so it will become more popular soon. Same as sports analytics, I anticipate growth in that field. I live in the middle east, and there are so many companies have initiatives 2030, 2050, zero emissions etc. So someone needs to analyse that data, create dashboards etc. Emissions are the problem, climate is a problem. I can imagine fast fashion will also do something around sustainability as look at all the waste they produce.

@@karinadatascientist thanks a lot for your response, Karina. Do you offer some one-on-one teaching or consultations?

@@nicatshiraliyev5793 yes, can you please reach out to my email karinadatascientist@gmail.com and we can take it from there :)

How exciting! Good luck with your studies, I hope some other videos (the project one) can be useful for you too

Thank you

Love it! Thank you for your comment. Also like the part where you say that real world project is an experience - 100%. If you use real-life datasets (not Titanic) to solve real life problem (Analyse data to see what stage of the funnel we see drop off; analyse marketing data to explain what is the most expensive ad and what is conversion; build dashboard) it is an experience similar to what you get at internship or other job. And if potential employer doesn't recognize it, then maybe it is not your employer. Another example, at my previous job, we needed to hire a person to cover maternity leave. We were looking for person with experience, so we can get them up to speed very quickly, but we looked not only at analytics experience, but at job experience. Someone who knows that with emails you "Reply All", can book a meeting, knows office culture. We had no time to train a person fresh from university. So sometimes it is the case, it is not about your programming/quering/dashboard skills, it is about "office cultural fit"

@@karinadatascientist well glad that you love it. Like you said there are some key factors which doesn't help with experience. I would like to know how does office environment works. If you have free time is it possible to connect at LinkedIn and talk when you have free time. Looking forward to learn more from you if you have some free time or if you want. I will be thankful to you! And best wishes for you.

unfortunately I don't know enough about data engineering, only high level things like data pipelines, Spark, Scala. Don't want to give misleading information

The jobs are unfortunately very very competitive now, not only in Asia. It got even worse because of AI hype, so at the interviews people start asking about AI chatbots and some crazy AI ideas, while you can help them today to increase revenue, fix funnels with data analysis, but it doesn't sound exciting, they want AI, facial recognition, flying cars, so then as managers they can add it to their resume to get paid more. I was recently asked at the interview if I do robotics (like literally, if I can build a physical robot). I think it's not too late to get into data field, it is a perfect timing. I know many people in finance (incl compliance, risk, internal audit) who lost jobs 6-12 months ago and can't find a job, or salary -40%. It's the market. Just accept that it might take 700 job applications to get an interview, look at it as a game. Don't look at rejections as it is something wrong with you. Also, you might need to work with recruiters to help you find a job. They can advise you on resume and they can recommend you directly to HR people in the companies.

@@karinadatascientist thank you for your perspective, may i ask u 1 more question which course or platform should i take to improve my data skills: dataquest, datacamp or 365datascience( if u took these course before) or there r some courses that u would recommend.

@@huyho5754 I am a product of W3 schools. I love the website, and when I was learning I was always looking up syntax there (now I use ChatGPT if I don't remember something). I've heard very good reviews about dataquest. As for courses - I think it is important who is the trainer, with some people you just don't like their way of teaching, so it's not going to work. I think most courses are ok, but stay away from General Assembly - too expensive, too rushed, no practical real life projects (at least it was like that sometime ago, maybe they changed their curriculum)

Glad it was helpful!

Thank you for watching

What a cool comment. Thank you!

That's a great idea, thank you

Thank you for watching!

Sure, very good idea

Thank you so much!

Sure. Links to the datasets in the description to the videos

Unfortunately it takes time to make videos, especially edit them. I wish I could post more often :)

Thank you! 👍

Great! I will think about a project in R

@@abdiwelly5606 I’ll try that! I am still learning how to edit my videos 😊

Sure, good idea!

This is the script I used: feature_importance = pd.DataFrame({'Feature': X.columns, 'Importance': model.coef_[0]}) feature_importance = feature_importance.sort_values(by='Importance', ascending=False) Can you try it? The error you get suggests that you’re trying to call model.coef_ as if it were a function, potentially you have extra (). For example, this is wrong coefficients = model.coef_() and this is wright coefficients = model.coef_

@@karinadatascientist It worked now. Thank you so very much

@@JYFrimpong Yay, glad to hear that

Thank you for watching!

Of course. You can download dataset here - www.kaggle.com/datasets/ahmedabbas757/coffee-sales/data

Thank you!

Yay! Glad you liked you and thank you for subscribing

Glad it was helpful! Thank you for watching

More to come! Thank you for watching

Привет 👋

You can use Linkedin or Indeed to find a job in the country you want. Find recruiters who advertise a job and send them DM

thank you! Glad you liked it

Awesome, thank you!

That's so cool! Hope you are enjoying it

thank you

Yes, you can

Very good question. The model doesn't actually "know" what "Male" or "Female" means. It treats encoded values as distinct identifiers and finds statistical relationships between the identifiers and the target labels during training.

@@karinadatascientist thank you for your amazing informations and answers. I have another one What if we have two columns we need to encode Sex column for example And health condition Good or bad Now we have two columns with 0,1 What is the process exactly in this case? How can we teach the model what is what?

The process is exactly the same. You will just have one column Gender with values 0, 1 Another column Health with values 0,1. Then model will learn patterns, and as a result might come up with a conclusion, that when Health is 1 and Gender 0, then this person has higher probability to do something. And then you need to interpret results back and say that health 1 is good, gender 0 is female, so female with good health has higher probability for something

Glad you liked it!

this is a good summary 😁

Thank you for watching

I agree. If you need to share code with team mates, if you need to deploy that model later, if you need to create a project for portfolio - good old fashioned method where you install all the libraries - is preferred. If you need to quickly analyse data for yourself - pyforest is pretty cool

Thank you for feedback. More projects to come!

Thank you for your support!

Thank you!