The libraries that Python needs to use in data science:
A. Numpy: Scientific Computing Library. A library that provides matrix operations.
B. Pandas: Data Analysis Processing Library
C. SCIPY: Numerical calculation library. The numerical integration and the solution algorithm of ordinary differential equations are provided. Provides a very broad set of specific functions.
D. Matplotlib: Data Visualization Library
E. Scikit-learn: Machine Learning Library
The installation sequence is as follows:
1.pip Install NumPy
2.pip Install Pandas
3.pip Install SciPy
(sudo apt-get install Libatlas-base-dev gfortran//This step is required to install scipy later.) Note: This I did not install first, directly installed scipy)
4.pip Install Matplotlib
(Install libpng before installing Matplotlib.) : Http://pan.baidu.com/s/1o80C5Jk, unzip into the folder and enter the Python install setup.py.
Install FREETYPE,:HTTP://PAN.BAIDU.COM/S/1DE5YVBR, unzip into the folder, enter Python install setup.py.
For FREETYPE.TAR.BZ2 package, install and unzip First, command: Yum install bzip2,bunzip2 FREETYPE-2.6.3.TAR.BZ2,TAR-XVF freetype-2.6.3.tar)
5.pip Install-u Scikit-learn
It is recommended to install with PIP. can also be installed with Yum, Apt-get installation. However, Pip installs a package that is generally newer, so it is recommended to use PIP for installation.
In fact numpy,pandas,scipy,matplotlib can directly download a anaconda on it.
Specifically, you can see http://python.jobbole.com/86236/.
Install Numpy,pandas,scipy,matplotlib,scikit-learn under Linux