您现在的位置：程式師世界 >> 編程語言 > >> 更多編程語言 >> Python

Top 10 Python libraries you should know in 2022

編輯：Python

Reading guide ： Let me introduce it to you , Most commonly used in data analysis 10 individual Python library .

author ：Terence Shin

compile ：Mika

source ：CDA Data Analyst （ID：cdacdacda）

Learning data analysis is by no means easy , There are countless tools and resources available . therefore , Sometimes it's hard for us to figure out what skills to learn , Which tool should I use .

In this paper , Let's introduce it to you —— Most commonly used in data analysis 10 individual Python library . Look at these libraries. Have you ever used them ？

01 Pandas

In the daily work of data analysts ,70% To 80% Both involve understanding and cleaning up data , That is, data exploration and data mining .

Pandas Mainly used for data analysis , This is the most commonly used Python One of the Libraries . It provides you with some of the most useful tools to explore data 、 Clean up and analyze . Use Pandas, You can load 、 Get ready 、 Operate and analyze all kinds of structured data .

02 NumPy

NumPy Mainly used to support N Dimension group . The robustness of these multidimensional arrays is Python List 50 times , It also makes NumPy Become the favorite of many data scientists .

NumPy By TensorFlow Other libraries are used for the internal calculation of tensors .NumPy Provides fast precompiled functions for numerical routines , These functions may be difficult to solve manually . For better efficiency ,NumPy Using array oriented computation , So you can easily handle multiple classes .

03 Scikit-learn

Scikit-learn Can be said to be Python The most important machine learning library in . In the use of Pandas or NumPy After cleaning and processing the data , Can pass Scikit-learn Used to build machine learning model , This is because Scikit-learn Contains a large number of tools for predictive modeling and Analysis .

Use Scikit-learn There are many advantages . such as , You can use Scikit-learn Build several types of machine learning models , Including supervised and unsupervised models , Cross validation of the accuracy of the model , Carry out feature importance analysis .

04 Gradio

Gradio It allows you to build and deploy a machine learning model in just three lines of code web Applications . Its use and Streamlight or Flask identical , But deploying the model is much faster , And it's much easier .

Gradio Its advantages lie in the following points ：

Allow further model validation . say concretely , Different inputs in the model can be tested interactively
Easy to demonstrate
Easy to implement and distribute , Anyone can access... Through public links web Applications .

05 TensorFlow

TensorFlow It is the most popular method for implementing neural networks Python One of the Libraries . It uses multidimensional arrays , Also known as tensor , Can perform multiple operations on specific inputs .

Because it is highly parallel in nature , Therefore, multiple neural networks and GPU To get an efficient and scalable model .TensorFlow This feature of is also called pipelining .

06 Keras

Keras It is mainly used to create a deep learning model , Especially neural networks . It is based on TensorFlow and Theano above , It can be used to simply build neural networks . But because of Keras Use the back-end infrastructure to generate calculation diagrams , So compared with other libraries , It's relatively slow .

07 SciPy

SciPy Mainly used for its scientific functions and from NumPy Derived mathematical functions . The library provides statistical functions 、 Optimization function and signal processing function . In order to solve differential equations and provide optimization , It includes the function of numerically calculating the integral .SciPy The advantage is that ：

Multidimensional image processing
Ability to solve Fourier transform and differential equations
Because of its optimization algorithm , It can calculate linear algebra very robustly and efficiently

08 Statsmodels

Statsmodels It is a library good at core Statistics . This multi-function library is a mixture of many Python The function of the library , For instance from Matplotlib Get graphic properties and functions in ; Data processing ; Use Pandas, Similar treatment R Formula ; Use Pasty, And based on NumPy and SciPy structure .

say concretely , It is important for creating OLS Such statistical models and performing statistical tests are very useful .

09 Plotly

Plotly It is absolutely a necessary tool for building Visualization , It's very powerful , Easy to use , And can interact with visualization .

And Plotly Also used with Dash, It can be used Plotly Visual tools for building dynamic dashboards .Dash Is based on web Of Python Interface , It solves this kind of analysis web In application JavaScript The needs of , And let you draw online and offline .

10 Seaborn

Seaborn Based on the Matplotlib On , Is a library that can create different visualizations .

Seaborn One of the most important functions is to create a visual effect of enlarged data . So as to highlight the relevant performance that is not obvious at first , Enable data workers to understand the model more correctly .

Seaborn There are also customizable themes and interfaces , It also provides data visualization with a sense of design , Better data reporting in .

Reference link ：

https://www.kdnuggets.com/2021/03/top-10-python-libraries-2021.html

Extended reading

Extended reading 《 utilize Python Data analysis 》

Dry goods go straight to

The development of Cloud Computing 4 Stages , At last someone made it clear
The three door problem of counter intuition ,80% All people are wrong ？
These new books of blockbuster , It's time to hoard
From consumer Internet to industrial Internet ： The wave of digital transformation is bringing about these changes

More exciting

Enter the following dialog box in the official account dialog box key word

See more quality content ！