What literature (Russian-language) would you recommend, according to which you can study data science?

Question

Books

I would like to know about the literature - preferably Russian-language -, according to which you can study data science (if any). I want to cover DS in a decent amount, so the books should be informative and complete (there can be a lot of them).
Thank you very much.

Answer 1

I'll give you an unpopular opinion. First, be sure to make sure that you are familiar with the basics of statistics and probability theory at the level of solving everyday problems. You can not climb into the wilds, but tasks from real life are worth clicking like nuts. Then make sure you know sql or nosql (if you absolutely need big data). And it will be absolutely great if you can understand how data is stored on a computer (for example, why a utf8 character can be 1 or 4 bytes).
Then yes, you can read books. They are great to apply in practice, but nowhere in the Russian Federation.

Answer 2

Machine Learning:
Theory - Peter Flach Machine Learning.
Practice - L. P. Coelho, V. Richard - Building machine learning systems in Python.
There are many articles on the Internet.

Answer 3

start with this one, then search
for example, https://habrahabr.ru/post/66561/

Answer 4

It is necessary to study the basis of mathematical analysis and higher algebra. Books contain too much practice and micro-recipes that are easy to work out on real projects. But theories are few and far between. Here's an example: I bought a book " Python and Data Analysis ". I can not say that it helped a lot, but in some ways it is really useful. All sorts of composite indexes, processing Excel files, visualization, manipulations with arrays ... But that's why in one case you need to take one method, and another another. or where it is necessary to discard extreme values ​​and where not - a mystery.
Later, I accidentally came across a crowdfunding project " Statistics and Cats"and bought this book. After reading, I got an understanding of the acceptability of certain methods of data analysis. I even manage to come up with these methods myself (what characteristics of the objects of analysis and how to measure, where to apply the system of weights and points ...). I believe that a paper book is already not to buy.But you can find an electronic one.I highly recommend it.
PS There is one more aspect in which it is necessary to understand for the successful analysis of the data. So to speak, the subject area, understanding how the digital characteristics of the objects of analysis are obtained. Without this, the analysis methods have to be selected by touch and not always correctly. Unfortunately, you don't always understand how things are measured. For example, when digitizing the characteristics of texts or images. There are a lot of examples on the Internet, they use ready-made libraries that give out matrices and vectors from text or pictures ... Then these matrices are folded and unfolded, vectors are multiplied, etc. But why and why in this order? To apply machine learning to all this, IMHO, it is very useful to understand how these matrices and vectors are obtained. Without this, you feel like a monkey who knows what manipulations to do, to get a banana (find similar texts or pictures), you can even understand how and why the banana moves along the conveyor to the cage, but you don’t understand where the banana comes from. :)