Catalog
1. When and how to use R？
2. When and how to use Python？
3. R The advantages of
4. R The advantages or disadvantages of
5. R Deficiency
6. Python The advantages of
7.Python Advantages or disadvantages of ： visualization
8.Python The shortcomings of
picture source ：https://www.datacamp.com/tutorial/r-or-python-for-data-analysis
To help developers learn and improve quickly , I opened up 《 FAQs for novices on the road 》 The column , Put your questions together , I hope I can give you some quick guidance , Avoid digging holes for yourself , Little detours .
About R and Python In the application of Data Science , Today, let's make a comparison .
If the data analysis task requires independent calculation or analysis of each server , It is recommended to use R. It is very suitable for research and exploratory work , It can be used for almost any type of data analysis , Because a large number of software packages and easy-to-use tests can usually provide the tools needed to start and run quickly . R It can also be applied in big data solutions .
R The following popular packages are available ：
dplyr,plyr and data.table Used to easily operate the package ,
stringr Processing strings ,
zoo For normal and irregular time series ,
ggvis,lattice and ggplot2 Visualization data ,
caret For machine learning
When data analysis tasks need to work with Web When application integration or statistical code needs to be merged into the production database , have access to Python. As a fully mature programming language , It is a good tool to realize production and use algorithms .
utilize NumPy / SciPy（ Scientific Computing ） and pandas（ Data processing ） package ,Python It can be used for data analysis ,matplotlib It can be used to draw pictures ,scikit-learn It is the application package of machine learning .
R There is a good Visualization Toolkit
Visual data is often more effective than raw numbers , It is also easier to understand . Visualization package ggplot2,ggvis,googleVis and rCharts All have very good functions .
R There is a good ecosystem
R It has a rich and cutting-edge package ecosystem and active communities . Can be found in CRAN,BioConductor and Github download R package , Can pass Rdocumentation Search all R package .
R Is the universal language of data science
R Developed by statisticians . They can go through R Code and software packages to exchange ideas and concepts , You don't have to have a background in computer science to get started . Besides ,R It is increasingly adopted outside academia .
R Appearance , Helped statisticians , But it increases the running time of the computer . Although due to the lack of code ,R It's very slow , But there are many packages that can improve R Performance of , Such as pqR,renjin and fastR,Riposte wait .
R It's not easy to learn , Specially , If from GUI Statistical analysis can be very difficult . If the R Not familiar with , Even finding packages can be time consuming .
have access to IPython Notebooks are easy to use Python And data .
It's easy to share notebooks with colleagues , Without having to install any programs , Can greatly reduce the organization code 、 The cost of output and note files , Can improve work efficiency .
Python Is a simple and intuitive universal language
Python Is a simple and intuitive universal language . It's easy to learn , It also improves the efficiency of developing programs . You can check my article 【 FAQs for novices on the road 】 About Python
Besides ,Python The test framework is a built-in and easy-to-use introductory test framework , Good test coverage , Code reusability and reliable performance are guaranteed .
Python There are some good visual Libraries , for example Seaborn,Bokeh and Pygal. Besides , And R comparison ,Python Visualization of is often more complex , The results of the demonstration are not ideal .
Python yes R Challenger . It doesn't offer hundreds of essential R An alternative to packages .
When doing data analysis , What's the use of R still Python Well , According to the above comparison , You should have some judgment .
