One of the main questions for people studying something is the choice of information sources: courses, books, articles. Recently, and especially in the study of data science, the breadth of choice is stupid: there are just a lot of courses and books, especially if we consider those available in English. In this article I will give a subjective overview of online learning platforms, showing those that provide more fundamental knowledge in the first place (spoiler: coursera.org and learning.oreilly.com are my favorites).
Introduction. Motivation for this article
When I started studying machine learning and mathematics, at some point I noticed how easily I understand those things that I learned rather than understood at the institute. This was due to several factors:
of course, the higher the motivation, because when you are engaged in self-education, you usually have a clearer idea of your desires and goals than at the university;
at every moment of time when something "does not go", you can postpone the current lecture / book and refresh / study the knowledge necessary for this topic; and good courses and books often start with the basics altogether, making sure that you understand all the prerequisites for each topic, i.e. have a foundation on which new knowledge can be consolidated.
since you can choose absolutely any course and teacher - you can find the one who explains the most clearly;
Finally, the huge benefits of video lectures are that you can pause learning if something just keeps you from learning. And you can listen to the lecturer several times, in contrast to a lecture at the university.
As a result, I got the opinion that the effective study of the material, first of all, depends on the quality and form of its presentation, and not on the student's talents. Therefore, for myself, I decided that it was worth spending energy on finding the best books and courses of their kind. As a result, while studying the basics of DS, I looked through dozens of different sources, and noticed that there are real gems that perfectly explain the material, and, on the contrary, many courses and books do not provide a fundamental understanding of the subject, but simply give individual examples, encouraging them to repeat ( type the code), but without explaining the general principles so that you can solve a fundamentally similar but different problem. Because the publication of books and courses is a profitable business, then there are much more courses and books of the second type:most of the books and courses that I looked through did not provide anything useful (memorable).
At first I started writing an article with recommendations for books and courses, but since new books and courses are published regularly, then it has changed that it is necessary to have an idea of online platforms in general. In addition, the platform overview turned out to be long, and it was worth publishing it separately from the list of specific courses. Below is a subjective comparison of platforms, mainly based on the courses that were available there in 2016-2018 (during these years I looked through dozens of options).
Platforms
coursera.org
This platform contains the best courses and specializations from what I have come across.
- , - . .. - - . , . (Google, , deeplearning.ai, ..)
( .. " data science") .
, - . , (+). - . , , - .
Andrew Ng (deeplearning.ai), , . "deeplearning.ai" .
coursera.org ( , ). - . "" , , ( https://www.coursera.org/hse).
, - .
coursera.org
Coursera.org . :
( ) " ". . , - , .). - / . , , .. , , .
- coursera.org , . . ( - ) - .
edx.org
coursera.org, . , .. coursera.org . , , -, coursera.org. , , , .
stepik.org
. / ( ). , , ( ). , , . , (?) , . , .
, ( coursera.org, . , ) .
, , , . , , . , , . ( ) .
udemy.com
, , . , . , Spark PySpark ( 2018).
, , , SQL ( ). PowerBI ( , ).
- . .
datacamp.com
, .. . , " " . , , ( Python, Pandas, SQL, ). , , udemy, .. datacamp data science, .
, , - . , , . .
, - . " " 50%.
: learning.oreilly.com ( )
, , , . , .
O'Reilly, , , , -. . , -.
. , , - >4 . , , . , .
- , . 2019, . , , , coursera.org, . , - .
, (~400$ ). oreilly.com - 2 ( ! - ). 50%
: b-ok.org
machine learning data science . , . . , .
, , data science - , , . , "! !" .
: amazon.com, ( ) ( ). 4 . 4 , , , . , , : 4.5+ , - ( ). , amazon . b-ok.org. , .
khanacademy.org -
( ) . - , .
, , (high school) coursera.org, , , . coursera.org " data science", . coursera.org , , / . , , .
: khan academy - ( ) , .
"" : ., skillfactory, geekbrains
( ?) , ., skillfactory, geekbrains. .
:
( )
, , coursera.org, ( ). , coursera.org , , , . - . , - , . , .
, , coursera . , , .
., skillfactory, geekbrains
- - (coursera.org) .
"" . , , .
- -. , , , ( .. ). , . - , . , , : - kaggle , , , coursera.org.
: ( - , ). , . - ,
?
- / - , , . , .
,
/
, , , .
, - self.development.mentor gmail.com,
, . . , (, ), / , , , - (?) .
- .