Which is better for statistical analysis, R or Python?

Python and R are the two most popular programming languages in statistics. The functionality of R is mainly considered by statisticians when developing (R has powerful visualization function), and Python is accepted by everyone because of its easy-to-understand syntax.

R Introduction Ross Ihaka and Robert Gentleman created the open source language R in S language in 1995, aiming at providing a better and more humanized language for data analysis, statistics and graphic modeling.

At first, R was mainly used in academic and research, but recently, the business community found that R is also very good. This makes R in Chinese one of the fastest growing statistical languages used by enterprises in the world.

The main advantage of R is that it has a huge community, supported by mailing lists, user-contributed documents and a very active stack overflow group. There is also CRAN image, a knowledge base containing R packages, which users can easily create. These packages are full of functions and data in R, and the images everywhere are backup files of R website, which are exactly the same. Users can choose the image closest to you to access the latest technologies and functions without developing from scratch.

If you are an experienced programmer, you may not think that using R can improve efficiency, but you may find that learning R often encounters bottlenecks. Fortunately, there are many resources now. Introduction to Python Python was founded by Guido van Rossem at 199 1, emphasizing the efficiency and readability of the code. Programmers who want to do in-depth data analysis or apply statistical techniques are the main users of Python.

The more you need to work in an engineering environment, the more you will like Python. It is a flexible language, which performs well in dealing with some new things and pays attention to readability and conciseness. Its learning curve is relatively low.

Similar to R, Python has packages, and pypi is the repository of Python packages, which contains many Python libraries written by others.

Python is also a big community, but it is a bit scattered because it is a common language. However, Python claims to be more dominant in data science: the expected growth and origin of newer scientific data applications are here. R and Python: Comparison of Numbers On the Internet, you can often see numbers that compare the popularity of R and Python. Although these figures often show how these two languages have developed in the overall ecosystem of computer science, it is difficult to compare them side by side. The main reason is that R is only used in the environment of data science, and Python, as a universal language, has a wide range of applications in many fields, such as network development. This often leads to the ranking results biased towards Python, and the wages of practitioners will be lower. How to use r? R is mainly used when the data analysis task requires independent calculation or analysis by a single server. This is exploratory work, because R has many packages and ready-made tests, which can provide the necessary tools to quickly start and run a large number of almost any type of data analysis. R can even be part of a big data solution.

When you start using R, you'd better install RStudiouide first. Then I suggest you take a look at the following popular packages:

Dplyr, plyr and data.table can easily operate packages? Stringr operation string? Does zoo do regular and irregular time series work? Ggvis, lattice and ggplot2 for data visualization? Caret machine learning

How to use Python?

If your data analysis task needs to use a Web application, or the statistical data of the code needs to be integrated into the production database, you can use python, as a fully mature programming language, which is a great tool to realize the algorithm.

Although python packages were still in the early stage of data analysis in the past, they have improved significantly over the years. NumPy/ SciPy (scientific calculation) and pandas (data processing) need to be installed in order to use Python for data analysis. Also watch matplotlib do graphics and scikit-learn machine learning.

Unlike R, Python has a clear and very good IDE. We suggest you look at Spyder and IPython websites to see which one suits you best. R and Python: The performance of the data science industry If you look at the recent opinion polls, R is the obvious winner in the programming language of data analysis. More and more people are turning from R&D to Python. In addition, more and more companies combine these two languages.

If you plan to work in the data industry, you should learn these two languages well. The recruitment trend shows that the demand for these two skills is increasing, and the salary is much higher than the average. R: advantages and disadvantages advantages: strong visualization ability. Visualization usually allows us to understand the numbers themselves more effectively. R and visualization are a perfect match. Some visualization software packages that must be seen are ggplot2, ggvis, googleVis and rCharts.

A perfect ecosystem R has active communities and rich ecosystems. R is wrapped by CRAN, Bioconductor and Github. You can search all R packages through Rdocumentation.

R of data science was developed by statisticians. They can exchange ideas and concepts through R codes and packages. No computer background is required. In addition, the business community is increasingly accepting R. The disadvantage is that R is slow and it is easier for statisticians, but your computer may be slow. Although R's experience is slow, there are several packages that can improve R's performance: pqR, Ren Jin, FastR, Riposte and so on.

R is not easy to learn deeply. R is not easy to learn, especially if you want to do statistical analysis from GUI. If you are not familiar with it, even finding the package can be very time-consuming. Python: Advantages and Disadvantages IPython NotebookIPython notebooks make it easier for us to use Python for data work. You can easily share notebooks with colleagues without them installing anything. This greatly reduces the cost of organizing code, output and annotation files. You can spend more time doing practical work.

Python is a universal language, simple and intuitive. It will be easier to learn, and it will speed up your programming. In addition, the Python test framework is built-in, which can ensure that your code is reusable and reliable.

Python, a universal language, brings people from different backgrounds together. As a universal and easily understood by most programmers, you can easily communicate with statisticians, and you can integrate each of your work partners with a simple tool. Defect visualization is an important criterion for selecting data analysis software. Although Python has some good visualization libraries, such as Seaborn, Bokeh and Pygal. But compared with R, the results presented are not always so pleasing to the eye.

Python is the challenger Python is the challenger of R, and it doesn't provide the necessary R package. Although it is catching up, it is not enough.

What should you learn in the end: the decision is up to you! As a data worker, you need to choose the language that best suits your work needs. Asking these questions before studying can help you:

What problem do you want to solve? What is the net cost of learning a language? What tools are commonly used in your field?

What other tools are available and how to make these commonly used tools?

Wish you success!

Excerpt from wechat article