Thank you so much, Danil. I watched it in one breath. You really know how to break down complex subjects into manageable, understandable pieces. I'm looking forward to following your suggestions. I'm very grateful for the privilege of knowing you personally and observing your growth and your attitude towards supporting others.
Thank you for compiling all the courses👍 Personally I'd appreciate more videos from you on getting started on Kaggle, it's quite discouraging at first tbh
Glad that it helped. I will release more on the topic. But in general: just grab any competition that is interesting for you, no matter what it is, and go for it. Then proceed to the next one. Learn from the published discussions and notebooks. Before you know it - you will start to understand what is going on. After getting some basic knowledge through courses, Kaggle will cement your skills.
@@lifecrunch I am currently pursuing a master's degree in data science at Durham University in the UK. After watching this video, I have decided to defer my course to next year and focus on Kaggle projects and internships instead. I was wondering if you have any Discord communities related to AI and data science that I could join.
Great plan. Regarding communities: - I have a following on medium where I write about DS practices medium.com/@danilzherebtsov - And you might want to check out my GitHub repository where I develop the library for Data Scicence tasks automation github.com/DanilZherebtsov/verstack
I know this is very relevant, but I somehow skipped this step and went directly for Data Science. There is a great video on this topic from another creator, you can check it out here ruclips.net/video/AYWLZ1lES6g/видео.html
Hi Danil, Thank you for your video. I do have a Statistics bachelor's degree from outside of the U.S. Do you think it might be slightly easier to land my first job?
Hi @Danil, I hope you are doing well. I just started my Data Science learning journey. I have a request, which course are you talking about (I think you are talking about this "Competitive Data Science course materials:" step 5) could you please upload those in a place from we can download them for our future use please. I have just started, but before jumping on it fully just wanted to gather all knowledge and resources in one place so that during study don't want to lose focus only to find things here and there. Thanks
This is a great video! Do you have any advice on learning how to build open source libraries like the one you made and whether you have to be an expert to build them, things like that would be great for us beginners!
Hey, thank you for this video, I'm currently focus on data visualization on Power BI and also plan to learn econometric (regression, time series,..) for my research but this subject is hard and dry, can you recommend me some project for this ? After that I will learn and do the kaggle challegne that you suggest. Thank you in advance
TimeSeries prediction in general can be tackled with three approaches: - simple and fast: autoregressive models like SARIMA, Holt Winters algorithm. Those even have some no code tools to facilitate TimeSeries prediction. This is a go to if the data has easily distinguishable trends. - moderate: using traditional ML approaches with correct validation, but here you might need some extra features, at least a time stamp. - complex and powerful: Sequence models, as I've mentioned in the description (imp.i384100.net/time_sereis_course) But in any case the data has to be prepared in order to run it through any of the mentioned algorithms. For small univariate time series you might get away with Excel or some BI Tools, but for complex data you will need to go with python or R. Try the Store Sales forecasting on Kaggle. The data is well prepared and you will have reference in the Discussions and Code tabs.
Python capabilities are enormous and I believe it is almost impossible to cover everything. The way to go about it is to study the basics and the other areas of this language will get revealed with practice. Sure you can study Python and Data Science together. In this case start from the third step on my list.
Very important question to be answered by you: What is the effective way to learn statistics and math for this field? I hope you will post a detailed video as so many guys looking for this answer from an experienced data scientist.
Well, the initial understanding will come from courses. The most valuable math coverage is in the linked set of 5 deep learning programs. The rest comes from practice/research. And there are a few great books, I will probably make a separate video on that later.
@@i_youtube_ In the video description there is a list of 5 Deep Learning courses (BEST NEURAL NETWORKS EDUCATION PROGRAM (in sequence). A lot of math is covered there.
I've just double-checked and the link works fine. This appears to be a Chinese web-site. Maybe you should try using some VPN service from your location.
what problem with me is that,,,,,,,,,,,,,,,,i am finding difficulty in just understanding data , when i have data after doing basic exploration .......what to do next? i think the problem is i am unable to interpret data. if you can help please let me know it'd be far better. thank you and obelized.
Data exploration should give you an idea about a few things: - which data types have you got and how to transform them (e.g. datetime column can be split into several derivative columns like 'Year', 'Month', 'Day', 'Holliday', etc. or a categorical column that is represented as an integer that should be transformed with mean-target-encoding...) - which columns should be removed as constant/noize/NaN/etc. - outliers that should be dealt with - missing values and how to better impute them - numeric data skewness and how to fix it - errors in data (e.g. a string in a numeric column) After having studied all of that you will have a clear picture on how to prepare your data for Machine Learning and after the data is prepared - it's a good place to select/choose/experiment a model type and fit a baseline model, record a score and proceed to improve it via: - different data transformation options - feature engineering - ensembling - model tuning
Hi, you spoke of a lady who used to work in a cafe, how relavent is past unrelated job experience to getting a job via this route? I work in a restaurant.
Data analysts and data scientists both work with data, the main difference is the tools they use and the function they execute. Data analysts examine data to identify trends, create dashboards and visual presentations to help businesses make strategic decisions. They work with databases and for interpreting the data mostly use no-code tools like excel/tableau/etc. Data scientists use programming language to interpret the data, visualize their findings and build predictive/descriptive model that will create predictions on new data. On a more advanced level Data Scientists may create ETL pipelines and deploy their models into the production environment.
Database interaction is an important part of machine learning. So yes, some sort of SQL dialect is going to be required in your future career. You may not possibly know which data warehousing technology will be your future employer's choice, but a basic level of SQL scripting should be acquired even before you need it. There are numerous online tools that help learn SQL and alike. Just google - practice SQL and give it some time to get familiar with the concept. Later, when reading a job description you are about to apply to, if you find the corresponding requirement, refresh it a few times before the job interview and you'll be fine.
Math knowledge level is really not that high. If you can handle such concepts as logarithm and maybe a derivative - you'll be fine. If you are completely new to math, it may be a bit overwhelming when going over the neural networks study, not because it's hard, just because there are a lot of simple operators that are combined in a large formula. But they still are just simple operators. And this is all going to be required during education. When working, you will be using mature libraries where all the underlying math is encapsulated, you will just be using documented APIs and methods that will run all the calculations for you. Although at some point in your education process it is very useful to code a simple neural network or a decision tree from scratch using nothing but numpy - this will give you an in depth knowledge and intuition about how machine learning algorithms learn.
@@lifecrunch Thank you a lot for the detailed answer. I am also wondering is math for Machine Learning engineer difficult or it's at the same level as in Data Science?
@@justrecorded4u984 Machine Learning engineers, apart from doing the actual Data Science work, execute models deployment, develop models monitoring and updating, etc. So the main difference between a Data Scientist and a Machine Learning engineer is not the depth of math, but the additional competences in software engineering, DevOps and production/cloud environment.
@@lifecrunch I very much appreciate your time and response. I'm studying Software Engineering and planning to transit to ML engineer later. I also was not bad at math at the university. That's my goal. Thank you again.
Thank you so much, Danil. I watched it in one breath. You really know how to break down complex subjects into manageable, understandable pieces. I'm looking forward to following your suggestions. I'm very grateful for the privilege of knowing you personally and observing your growth and your attitude towards supporting others.
Thanks for the kind words!
Thank you for compiling all the courses👍 Personally I'd appreciate more videos from you on getting started on Kaggle, it's quite discouraging at first tbh
Glad that it helped. I will release more on the topic. But in general: just grab any competition that is interesting for you, no matter what it is, and go for it. Then proceed to the next one. Learn from the published discussions and notebooks. Before you know it - you will start to understand what is going on. After getting some basic knowledge through courses, Kaggle will cement your skills.
Thanks for this video! Very interesting, looking forward to new videos.
Thanks for watching!
Thank you it's a great video and great start for your channel.
Thanks for the comment! Working on more cool content!
Wow! So much useful information! Thank you!🙏🏻
Glad it was helpful!
Too motivational to ignore!!! Keep on doing that crazy stuff!!!💪👌
That's the plan!
Fantastic video! Thank you so much for posting - I could not hit that subscribe button fast enough!
...music for my 👂 😏
If you upload high-quality videos like this, this channel will explode, trust me. Thank you ever so much.
Working on a great new video right now
@@lifecrunch I am currently pursuing a master's degree in data science at Durham University in the UK. After watching this video, I have decided to defer my course to next year and focus on Kaggle projects and internships instead. I was wondering if you have any Discord communities related to AI and data science that I could join.
Great plan. Regarding communities:
- I have a following on medium where I write about DS practices medium.com/@danilzherebtsov
- And you might want to check out my GitHub repository where I develop the library for Data Scicence tasks automation github.com/DanilZherebtsov/verstack
Thanks for guidance 🙇
Glad it helps!
Do a video on quickest way to get into data analytics
Similar like data science
I know this is very relevant, but I somehow skipped this step and went directly for Data Science. There is a great video on this topic from another creator, you can check it out here ruclips.net/video/AYWLZ1lES6g/видео.html
Thank you Sir, You're video was really helpful especially the course compilation. Looking forward for more guidance from you.
Thank you for the kind words, more content to come
I just hit that sub button. Thanks for this video.
Welcome ,) stay tuned!
You are amazing! Thank you for all this
Thanks for the good vibes!
great....kindly make a video regarding data analysts role as well
Unfortunately my competence in the data analyst track is limited. I have skipped this step and went straight into Data Science from the start.
Thank you Danil! this honestly has got me hyped as I didn't know where to start but it seems you have answered my question.
Great motivation for the start. Keep it up!
can you do a video about the best ds projects to do to impress a recruiter
That’s a great idea! Stay tuned.
Thankyou bro! Very Useful 👍
Welcome 👍
This is really helpful. I nearly lose interest but i found this video.......i'm gonna continue my learning ...big thanks!!!!!!!!
Motivation is 50% success ,)
Thank you for this mate! Need this one badly. I am a computer science student wanted to specialize in Data Science Learning
CS background is a great foundation for DS&ML!
great content! please make some more about the DATA an/sc topics
Thanks! What do you mean by an/sc?
ohh my god!!!! this video is a GEM!!!
🙈
I subscribed to help you get closer to 4000. You'll get there before you know it.
Thanks man!
I wanna know more Data science!!!
More to come
Hi Danil, Thank you for your video. I do have a Statistics bachelor's degree from outside of the U.S. Do you think it might be slightly easier to land my first job?
Absolutely. Not only the education is relevant to the DS role, but also it will be easier for you to learn the subject.
Hi @Danil, I hope you are doing well. I just started my Data Science learning journey. I have a request, which course are you talking about (I think you are talking about this "Competitive Data Science course materials:" step 5) could you please upload those in a place from we can download them for our future use please. I have just started, but before jumping on it fully just wanted to gather all knowledge and resources in one place so that during study don't want to lose focus only to find things here and there. Thanks
Unfortunately there is no way to download them. The platform which hosts them does not have the download option
Some of the links in the Description are not clickable can you fix it
Thanks for the heads up. Fixed it
This is a great video! Do you have any advice on learning how to build open source libraries like the one you made and whether you have to be an expert to build them, things like that would be great for us beginners!
This is a good idea for a separate video! I think I might just make it one of the next publications.
Hey, thank you for this video, I'm currently focus on data visualization on Power BI and also plan to learn econometric (regression, time series,..) for my research but this subject is hard and dry, can you recommend me some project for this ? After that I will learn and do the kaggle challegne that you suggest.
Thank you in advance
TimeSeries prediction in general can be tackled with three approaches:
- simple and fast: autoregressive models like SARIMA, Holt Winters algorithm. Those even have some no code tools to facilitate TimeSeries prediction. This is a go to if the data has easily distinguishable trends.
- moderate: using traditional ML approaches with correct validation, but here you might need some extra features, at least a time stamp.
- complex and powerful: Sequence models, as I've mentioned in the description (imp.i384100.net/time_sereis_course)
But in any case the data has to be prepared in order to run it through any of the mentioned algorithms. For small univariate time series you might get away with Excel or some BI Tools, but for complex data you will need to go with python or R.
Try the Store Sales forecasting on Kaggle. The data is well prepared and you will have reference in the Discussions and Code tabs.
@@lifecrunch thank you for your suggestion 😊
To me watching the bullet train. Need to watch more times.
It's got all you need broken down into steps. Just go one step at a time and you'll get there.
@@lifecrunch yes, my journey started. Thank you for your kind support 😊.
How much to learn python it is vast
and can we learn python and data science together
Python capabilities are enormous and I believe it is almost impossible to cover everything. The way to go about it is to study the basics and the other areas of this language will get revealed with practice.
Sure you can study Python and Data Science together. In this case start from the third step on my list.
Very important question to be answered by you: What is the effective way to learn statistics and math for this field? I hope you will post a detailed video as so many guys looking for this answer from an experienced data scientist.
Well, the initial understanding will come from courses. The most valuable math coverage is in the linked set of 5 deep learning programs. The rest comes from practice/research. And there are a few great books, I will probably make a separate video on that later.
@@lifecrunch sorry I didn't get it! What do you mean by those 5 programs?
@@i_youtube_ In the video description there is a list of 5 Deep Learning courses (BEST NEURAL NETWORKS EDUCATION PROGRAM (in sequence). A lot of math is covered there.
Does anyone knows how to open that bilibili link he wrote? I am not able to see the video..
I've just double-checked and the link works fine. This appears to be a Chinese web-site. Maybe you should try using some VPN service from your location.
what problem with me is that,,,,,,,,,,,,,,,,i am finding difficulty in just understanding data ,
when i have data after doing basic exploration .......what to do next?
i think the problem is i am unable to interpret data.
if you can help please let me know it'd be far better.
thank you and obelized.
Data exploration should give you an idea about a few things:
- which data types have you got and how to transform them (e.g. datetime column can be split into several derivative columns like 'Year', 'Month', 'Day', 'Holliday', etc. or a categorical column that is represented as an integer that should be transformed with mean-target-encoding...)
- which columns should be removed as constant/noize/NaN/etc.
- outliers that should be dealt with
- missing values and how to better impute them
- numeric data skewness and how to fix it
- errors in data (e.g. a string in a numeric column)
After having studied all of that you will have a clear picture on how to prepare your data for Machine Learning and after the data is prepared - it's a good place to select/choose/experiment a model type and fit a baseline model, record a score and proceed to improve it via:
- different data transformation options
- feature engineering
- ensembling
- model tuning
@@lifecrunch can you make video like how deal with missing values(e.g. how to impute them ) for redundant data the same.
@@imitative_entreperneur That is a great topic. I will see what I can doo.
Brilliant content watch in 1.25x it would be more interesting
Great comprehension ,) I've had viewers saying - 'it is too fast'
If one can implement Transformer from scratch, could that help in landing a job?
If you can explain how you did it to someone without a technical background, so that they would understand, then most definitely!
hey I'm struggling to find Python practice questions. Can you share some good practice questions for python.
Well the best one hands down is leetcode.com
Hi, you spoke of a lady who used to work in a cafe, how relavent is past unrelated job experience to getting a job via this route? I work in a restaurant.
I don’t even know what was her past experience. And it doesn’t matter. If you want something - just go for it.
@@lifecrunch thank you, do you think Power BI or tableau and SQL are worth learning too?
@@willroberts1891 these are very well in demand for a Data Analyst position, less so for a Data Scientist role, but they are not uncommon.
Данила, экселент эксент энд I think I should give a subscription to you)
Удачи ;)
Welcome)
Is data science and data analyst the same?
Data analysts and data scientists both work with data, the main difference is the tools they use and the function they execute.
Data analysts examine data to identify trends, create dashboards and visual presentations to help businesses make strategic decisions. They work with databases and for interpreting the data mostly use no-code tools like excel/tableau/etc.
Data scientists use programming language to interpret the data, visualize their findings and build predictive/descriptive model that will create predictions on new data. On a more advanced level Data Scientists may create ETL pipelines and deploy their models into the production environment.
@@lifecrunch gotcha , thanks for explaining and your video was also very helpful
Hi danil, when will you publish your videos related to data science stuff..like the teaching the things in your own way..
A video on how to create and publish your own python open-source library is in production. Stay tuned.
Thank you! Would you recommend learning SQL? If so, what level of SQL learning is required?
Database interaction is an important part of machine learning. So yes, some sort of SQL dialect is going to be required in your future career. You may not possibly know which data warehousing technology will be your future employer's choice, but a basic level of SQL scripting should be acquired even before you need it. There are numerous online tools that help learn SQL and alike. Just google - practice SQL and give it some time to get familiar with the concept. Later, when reading a job description you are about to apply to, if you find the corresponding requirement, refresh it a few times before the job interview and you'll be fine.
what about necessary knowledge in math?
Math knowledge level is really not that high. If you can handle such concepts as logarithm and maybe a derivative - you'll be fine. If you are completely new to math, it may be a bit overwhelming when going over the neural networks study, not because it's hard, just because there are a lot of simple operators that are combined in a large formula. But they still are just simple operators.
And this is all going to be required during education. When working, you will be using mature libraries where all the underlying math is encapsulated, you will just be using documented APIs and methods that will run all the calculations for you.
Although at some point in your education process it is very useful to code a simple neural network or a decision tree from scratch using nothing but numpy - this will give you an in depth knowledge and intuition about how machine learning algorithms learn.
@@lifecrunch Thank you a lot for the detailed answer. I am also wondering is math for Machine Learning engineer difficult or it's at the same level as in Data Science?
@@justrecorded4u984 Machine Learning engineers, apart from doing the actual Data Science work, execute models deployment, develop models monitoring and updating, etc. So the main difference between a Data Scientist and a Machine Learning engineer is not the depth of math, but the additional competences in software engineering, DevOps and production/cloud environment.
@@lifecrunch I very much appreciate your time and response. I'm studying Software Engineering and planning to transit to ML engineer later. I also was not bad at math at the university. That's my goal. Thank you again.
@@justrecorded4u984 Now you have all the materials you will need. Best of luck!
One day...
You will become big RUclipsr
best wish so far ,)
I wonder, what made you come to this conclusion?
Compressed and then stomach🤣🤣🤣
A little humor won’t hurt a serious subject 😉