Introduction to Data Science from an amateur

This material is not a competitor to courses in Data Science, machine learning, neural networks, artificial intelligence and other hype areas, because professionals teach there, at least it should be like this, but here the typist and word compiler is the same beginner. However, there is no better way to learn any subject or discipline than to try to teach someone else. This is the main reason why I am writing this series of articles. I did not come up with this method, but it really works, and not only in the direction of the teacher, but also in the direction of the student. Because it is very difficult for a student to break through to the transcendental thoughts of the guru, it is much easier to ask the same student: “Did you understand what he said?”

How else will this material differ from many courses? While classical education begins with a boring theory that students do not understand, not because they do not like it, or are stupid, or are not prepared to perceive this information, but because they do not know what to do with it, in which compartment to put these knowledge. And by the time that theory is needed, unfortunately, there is little, if any, left of it. Here we will first encounter a problem “nose to nose” in practice, and only then we will make references to the theory, and maybe even try to find it out. This positively motivates people to understand the subject, and maybe even get carried away by it. Even if this material helps only one person, then it will not be useless.

Let’s start with setting the problem that we will solve. Why not take a swing at our ruble, or rather the dollar, or rather the dollar-ruble currency pair, or, in short, the dollar-ruble. Recent events have shown that this financial instrument actively affects the lives of not only traders, banks, funds, etc., but also ordinary citizens, even if these citizens are not engaged in foreign economic activity. Therefore, it is very useful to have a forecast of this financial asset, even if this forecast is bad, but it is still better than the absence of it.

Here we simply have to make a clarification, a reminder, a warning that this material is not an investment recommendation, because we are making a forecast for educational purposes. And if you want to use this knowledge for selfish purposes, for example, for trading on the stock exchange, you will do it at your own peril and risk.

And the free, still available, web resource will help us with this. Google Collaboratory. There are other “enemy” similar resources, for example, Google has a whole platform for Kaggle competitions. The “native” Yandex has a similar free DataSphere cloud service, but its speed clearly hints at the use of a paid option. You can, of course, arrange your own laboratory in which to raise Jupyter Notebookbut we will use colab.

If for some reason you do not have any Google account, for example, in YouTube, then it’s time to get one. I think that with this you will not have any questions. Then you can go to the above link (just in case, I will duplicate it: colab. To create a new notebook (this is how a page or project is called here), you just need to select the “Create Notebook” item in the “File” menu.

We can immediately rename, for example, to FirstNote: File -> Rename. Let’s do one more action in colab, namely, connect the Google drive. This is necessary to save data between sessions.

Saving happens automatically, but before closing it is better to make sure that the last changes are saved. Now we are ready for the first lesson. I hope that the material was clear and useful.