Python for data analysis: Related Package installation, pythonpackage

Python for data analysis: Related Package installation, pythonpackage

1. Why use Python for data analysis?

Python has a huge and active scientific computing community with improved libraries that can easily integrate C, C ++, and Fortran code (Cython project ), it can be used for both research and prototype construction and production system construction.

Ii. Advantages and Disadvantages of Python:

1. Python is an interpreted language and runs slowly than compiled data. 2. Because python has a global interpreter lock (GIL) that prevents the interpreter from executing multiple python bytecode at the same time, python is not applicable to highly concurrent and multi-threaded applications. 3. Common extension packages for data analysis using Python.At present, the initial phase of learning involves the installation of four packages: numpy, scipy, pandas, and matplotlib. The Python2.7 version is installed in my notebook, and pip and setuptools are installed, for installation of pip and setuptools, see related notes. The initial installation command is simple:

pip install pandaspip install numpypip install scipypip install matplotlib

However, only numpy and matplotlib packages have been installed successfully. pandas and scipy have failed to be installed. After checking the relevant information, we found that it may be a version issue or the dependency of the package.

Finally, a great Python package was found in stack overflow. The URL: http://www.lfd.uci.edu /~ Gohlke/pythonlibs/# scipy

-- Mark it here and try to write a crawler to prevent loss of all the packages.

The above Web site is provided by the University of California, Owen, Python-related libraries, modify # The name behind can go to the download page of other packages, this page provides instructions on the front-end packages required to install a package, which is very friendly.

The dependency packages are described as follows:

Pandas, a cross-section and time series data analysis toolkit.Requires numpy, dateutil, pytz, setuptools, and optionally numexpr, bottleneck, scipy, matplotlib, pytables, lxml, xarray, blosc, backports.lzma, statsmodels, sqlalchemy and other dependencies.

 

Then there is a bunch of pandas.

Finally, the numpy + mkl whl file is installed based on the relevance of each package, then scipy is installed, and pandas is finally installed.

The installation method is as follows:

1. download the corresponding 4 packages in the D: \ directory (it is strange that my laptop is AMD64-bit, but the package installed with amd64 reports an unsupported platform error, 32-bit installation can be properly imported)

2. Run the cmd command line to enter the D: \ directory and run: pip install <full name of the package>. (If another version is installed, use pip uninstall to uninstall it)

Finally, run the following command to view the package installation location: