Transcription
Hello everyone. Today I will share my experience on how to quickly learn to program in Python for data analysis, even if you are currently a complete beginner and don't consider yourself to be in this field. I made this video solely to share my experience and to help those who currently think that starting to program in Python is quite difficult. Now I am sure that even a complete beginner can learn to program and do simple data analysis in just 2 weeks. Data analysis is not millions of lines of code; many data manipulations are just one or a few lines of code. It is important to note that this video is more for those who want to learn to program in Python for data analysis independently, so all video materials will be free. So, let's start with the first, important part: learn the basics of the Python language. What is included in the basics? In all paid and free courses, and I have watched many of them, the basics are approximately the same: working with different data types, for and while loops, working with functions, lists, and so on. But here a problem may arise: you study the basics for weeks, months, but still don't understand what a data analyst actually does. The syntax is important, but it is also important not to drown in it. For example, at the start, I studied a free course in two parts on Stepik. Is this course good? Yes, and I have recommended it myself more than once, but studying it takes quite a long time. In the course itself, for example, it is stated that the training in the first part lasts only 40 hours, but in my opinion, this is not the case. I definitely studied it for longer, in the first part alone, more than 150 tasks, and some tasks can take 2 hours to solve. Thus, you can spend several months studying the basics but still not understand how to do data analysis using Python. Here I will emphasize again: syntax is important, but the main thing is not to drown in it, and you will return to the syntax regularly anyway, delving deeper into it. Regarding short courses, Kaggle has a free course. All links are in my Telegram channel, attached at the bottom of the description.
The second point: learn libraries for data analysis. It is by studying libraries that you will understand how to do data analysis using Python. Libraries are a kind of code template for each specific situation. It is best to study libraries not in isolation, but by doing a practical project. The best way to start is with the Pandas library; it is the main library for data analysis. With it, you can clean data, prepare it, analyze it, and visualize it. As I mentioned at the beginning, many data manipulations are just one or a few lines of code. Let's look at an example with the Pandas library and see for ourselves. For example, to read a file, only one line of code is enough: `pandas.read_csv()`. If we want to find, for example, the average age of Titanic passengers, we also use one line. Or, for example, grouping data by category. And this is how we can create a graph. A link to a short, free course from Kaggle on the Pandas library will also be in the description. Pandas also has an official website where there are instructions on how to work with this library. In general, each library has its own website where there is a full description. If you already know, for example, Excel or SQL, then it will be even easier for you because Pandas has similar functions, and Pandas even has its own guide where Pandas functions are compared with Excel and SQL functions for better understanding. After mastering the basics of Pandas, move on to other libraries for data analysis. In principle, there are not many besides Pandas; there are also NumPy, Matplotlib, and Seaborn. There are also other libraries, but this is already, so to speak, at a higher level.
The next point: start a data analysis project from the first days. In general, learning through practice is the best way to learn to program. Yes, it is a challenge and a step out of your comfort zone, but you will learn much faster this way. After a short syntax course, you can already move on to a real project, simultaneously learning libraries. The main problem at this stage is not understanding what these data analysis projects look like and where to find them. The banal and general answer is the internet, but sometimes it's hard to google when you don't understand anything yet. Especially in the English-speaking internet, there are many such projects; they are posted by ordinary data analysts who simply share and explain how they do it in detail, with code, with all explanations, and so on. Especially many such projects are on Medium. I have made a special selection of several projects for beginners, and again, there are all the details where everything is explained and shown, how and what to do. True, they are on the English-speaking Medium, but if you don't know English, use a translator, for example, a Chrome extension. The link to these projects will also be in the description. Initially, look at these projects and do something similar with your own data. With what data? With any data that you might have, or if not, you can download it from the internet from sites like Kaggle; there are many datasets there, or also from UCI; there is also a special site with datasets. In the process of doing your first project, and not only the first ones, you will encounter various errors in the code. To solve these errors, you can google them, you can turn to Stack Overflow, or you can ask ChatGPT. Now we have ChatGPT here, and in principle, I have already resorted to this method; it turns out to be even more effective than turning to Google or Stack Overflow.
And the last point: to do analysis, to write code for analysis, you need some development environment. There are many different ones, but usually they need to be configured. In principle, the task is not so difficult to configure a development environment, but it took quite a lot of time at the beginning. Therefore, from the first steps, to not complicate your path, you can use cloud development environments, for example, Google Colab or Kaggle Notebooks. A cloud development environment is a code editor where you go and can immediately write code without any configuration. And in the future, you can set up a professional development environment such as Visual Studio Code, for example, or some other. Or you can set it up right away using ChatGPT instructions. This way will possibly be faster. Thank you all for watching. I have to say that if you liked the video, please give it a like, simply because YouTube says the more likes, the more your video is shared, the better it is promoted. [music]