NumPy, SciPy, Pandas, and Matplotlib are fundamental scientific computing and visualization packages with Python. Scikitlearn is a simple and efficient package for data mining and analysis in Python. NLTK (the Natural Language Toolkit) is a leading platform for building Python programs to work with human language data. However, if you are using Python on Windows, it seems not easy to install these packages. Thanks to Christoph Gohlke, who provides the nonofficial Python extension packages for Windows. With these unofficial packages, there is a simple way to install these libraries on Windows.
For clarity, I’m using Python 3.4.3 on 64bit Windows 7. The steps described below would also work for 32bit. Please note, after installed Python, I added the ‘Scripts’ directory in Python package where the ‘pip.exe’ and ‘pip3.exe’ located, in my case the path is ‘C:\Python34\Scripts’, to the PATH environment variables. This step makes my life easier, because I don’t need to navigate to the scripts directory anytime I run ‘pip’ or ‘python’.
Install NumPy, SciPy, Pandas, and Matplotlib
The steps of installation of Numpy, SciPy, Pandas, and Matplotlib are the same. Here I take NumPy as an example:

Download appropriate wheel file, in my case, for numpy, it is “numpy1.10.1+mklcp34nonewin_amd64.whl” from the unofficial extension packages site

On the command line, navigate to the directory where you saved the wheel file

Run
pip install numpy1.10.1+mklcp34nonewin_amd64.whl

To verify, start python and run
import numpy
numpy.__version__
Install Scikitlearn
With a working installation of Numpy and Scipy, it is easy to install scikitlearn through running: pip install U scikitlearn
Install NLTK
To install NLTK we need to install the Pyyaml package. Follow the steps for installing Numpy above to install Pyyaml first and then NLTK.