Is Python hard to learn? It's not difficult. I learned while working on a project. After half a year, I passed the interview of the editor of the publishing house, received a Python topic selection, and published it successfully.
Some students will say that you have the opportunity to practice programming based take-out projects, so learn fast. This is true, my previous foundation has indeed accelerated my learning progress. But having said that, I usually do part-time training. Most of my students have no foundation. I use the information and methods given in this article to train them. Classmates who have learned well will be able to do basic python projects in 3 months. At this speed, it should not be difficult to master python in one year and reach the level of publishing a book.
The so-called gifting rose hand has a lingering fragrance. In this article, I will restore my learning process in detail as much as possible, as well as the steps I tutored students to learn python, in order to maximize the help to those students who want to learn python.
1 To be clear about the purpose, but to learn other things if there is no purpose
If it is for learning to play, then it is not recommended to learn python, because after a few weeks of learning, you will lose interest and the previous learning will be wasted.
Learning Python generally requires a more utilitarian purpose. For example, if you want to learn to change jobs and increase your salary, you can use it for a company project like me, learn to earn wages, or use it for private work. After you learn it, you can record videos and publish books.
After you have a clear goal, you will have a learning direction. For example, if you have learned well and want to change jobs, you can check the requirements of most jd for python, which is nothing more than data analysis and crawlers with various machine learning and deep learning applications. For example, if I want to do financial quantification, it is nothing more than the Three Musketeers of data analysis and deep learning. I may also bring some crawler technology. If you want to publish a book or a video, you should also see how others do it. In short, once you have a utilitarian purpose, you will understand the list of content you should learn.
In fact, the bottom line is that even the deep learning knowledge graph artificial neural network seems to have very deep skill points, but in python it is nothing more than introducing a library + calling method, it is nothing more than passing in different parameters when calling, and these skills are all There are ready-made books for reference. But if you don’t know your learning goals in advance, you won’t know what else to learn.
I saw many people on the Internet asking, how to learn python? What should I learn? How do others know this kind of thing? In fact, you should ask yourself. You should learn what you want to make money with Python in the future.
2 You must first set up the environment and install the third-party package
A lot of nonsense, no matter what skills you want to learn in python, the development environment must be built.
1 Go to the official website https://www.python.org/downloads/windows/ to download the Python interpreter.
2 Download the integrated development environment, such as eclipse+pydev, or use Pycharm directly, I am using the latter.
3 To download the third-party package, the basic command is pip install, and you need to change the source if necessary.
What is a good Python development environment?
1 Can run the hello world program through python in integrated development environment such as Pycharm.
2 The third-party package can be downloaded correctly, here is numpy as an example, download this library, and can develop the numpy basic program in pycharm, that's fine.
When installing a third-party package, you may encounter the problem of version mismatch. The specific method is to put the error message on the Internet when you encounter an installation error, and you can always find a lot of solutions, and then solve it slowly . But this kind of problem is not often encountered.
3 What should I learn basic grammar?
I used the following two books for getting started, the first one is this.
It took me about two weeks to type the code of this book. Then type the code for the second book. Although this book talks about data analysis, it also talks about the basic syntax of python.
When I typed the codes of the two books, I felt that I could understand the basic python programs. When I wrote two Python books myself, I found that when I was getting started, I still took a detour. It was not that I misunderstood the knowledge points, but learned what I didn't need to learn at the beginner stage.
Now for reflection, after setting up the python development environment, you can use the following methods to get started with python syntax.
1 Just type some code such as if branch and while loop, understand the basic structure of python, and understand the practice of "indentation".
2 Need to look at basic data structures such as lists and dictionaries, because python is mainly used to analyze data, and these objects are used to store data when analyzing data.
3 Then look at the code in the exception handling part.
At the same time, the following code does not need to be looked at.
1 The code of object-oriented thinking, such as classes and inheritance, is basically not used in real projects.
2 The code of the file reading and writing part. Because the DataFrame used for data analysis can easily interact with files in csv and other formats, the method of reading and writing files that comes with the python core library is basically not used.
Other codes, such as the analysis of this article, regular expressions, and urllib, are not unimportant, but can be viewed in the specific context of the project, instead of simply looking at the relevant grammar like memorizing words, so that the efficiency of learning will be Very high.
4 What skills should be learned in data analysis?
At the moment, Python is most likely to be used for data analysis. General data analysis projects also come with some machine learning. When learning this part, my suggestion is: first go through the full stack process, and then look at the relevant methods.
First, data analysis is nothing more than using the Pandas library to read the data from the csv, then use the numpy method to analyze and clean, and then use the matplotlib library to draw the graph. You can pass this process through several cases.
I feel that this book I wrote was written around this idea.
Second, after the full-stack process is completed, you can look at a few points. The first is how to read and write csv (or other files) in the pandas library, the second is how to manipulate the row and column data in the dataframe object, and the third is how to use matplotlib. Library draws histogram pie chart and so on.
When learning, you don’t need to look at the grammar too closely, because after a few weeks you will forget the paintings you don’t use. When the project is actually done, you can find the corresponding methods and parameters according to your needs.
For example, I need to draw the stock data that I read into a line chart, and the line chart has requirements for scale labels and titles, then I need to check ready-made cases and corresponding apis based on this need, so that I can apply what I have learned immediately.
5 How to learn machine learning and deep learning?
When learning this part, be sure not to look at the algorithm first. For example, you must not pay attention to algorithms such as regularization, gradient descent, and dimensionality reduction, because they are not necessary and these algorithms are very deep. If you really look at the algorithms, it is estimated that your learning confidence will soon be destroyed.
How should I learn? Use the built-in data set and example study.
The sklearn library that encapsulates machine learning comes with many examples, such as the Boston housing price case describing linear regression, the Iris case describing SVM classifier, and other algorithms such as ridge regression. In this book, many cases are used to describe machine learning.
The same goes for deep learning, don't look at the algorithm, first look at the api. After you run the code of several books, you will have a good idea. And when everyone will find that the code to be written in your project is similar to the code framework you have read and typed before.
1 First, it is nothing more than creating various models, such as artificial neural networks, SVM, etc. When creating different models, it may be that the constructors and class names are different.
2 Then pass in the characteristic value and target value. For example, when I am doing stock quantification, the characteristic value is the highest opening price and the lowest trading volume, etc., and the characteristic value is the closing price. In different projects, as long as there are business requirements, the eigenvalues and target values must be determined. What we programmers have to do is to pass in the eigenvalues and target values in the fit method, and then let the model train.
3 After training, call the predict method to predict.
When working on a project, the predicted result may be different from the expected result. It doesn't matter, just wrestle slowly and just change in the parameter eigenvalues, or discuss the model together. In short, these are not technical issues.
6 I summarized a better learning method
First, you must set up a good environment, run the example code first, and then look at the code and methods through the effect, so that you can basically understand the key points after running the code.
Second, at the beginning, you must not go to the Internet to read the information, because the online information may not be a system first, and the second may not be able to run through. At the beginning, you must systematically type the code according to the book.
Third, when looking for introductory books, you can read books that teach basic grammar, but one is enough, and you can even read comprehensive books. For example, the several books I mentioned above comprehensively include grammar + data analysis + machine learning and other skills. This way, after understanding the grammar, you can directly enter the specific project skills link.
Fourth, when you first started a project, in addition to writing code yourself, you can also refer to other people's similar cases. For example, if I want to draw a stock K-line, and even draw a legend on the k-line, then I can find two or three people's similar codes, first run it, and then learn from others' methods and parameters, so that it is better than anything else. Want a lot better.
7 How to learn Python well?
Back to the topic mentioned at the beginning of this article: the most direct and effective way to learn how to realize Python is to change jobs.
The specific method is to take the master and the slave. For example, if your major is Java or big data, you still have to focus on this direction, but you should write in your resume that you are familiar with data analysis and machine learning, etc., and you are also familiar with neural networks and other models, and then you need to invest in such needs company of.
At present, companies that can use python for data analysis and machine learning are not necessarily low. So if you have both python and other main direction skills, it is not a big problem to enter such a company. In the training school, I often help students who learn python to do this.
Of course, you can also do sidework with python. For example, I have published books, recorded video lessons, and made teaching materials for colleges and universities. In short, I still have this sentence: first specify the utilitarian learning goals, and then refine the specific content to be learned, and then Learn systematically by reading books and typing codes. After you learn it well, you will definitely be able to make a profit.
8 Send e-books at the end of the article
This time I want to give you this book: "Practice for Getting Started with Python Based on Stock Big Data Analysis (Video Teaching Edition)", Jingdong link: https://item.jd.com/12868774.html.
I have done similar activities before, and if some classmates have already got it before, I won’t use it this time. I will give another python book I wrote later when it is appropriate.
The specific method is: please pay attention to my official account first: make progress together and make money together.
Then you can add my WeChat and indicate that you want a python stock book, so that I will give you the word electronic version of this book.
