Install NumPy on windows (with detail steps)

907

NumPy on your windows

Now we are going to discuss how to download and install the NumPy on your windows. The foremost thing you must have to install the NumPy is that you should download and install Python on your windows. In case if you do not have Python Installed then here is a short description to install it in your windows. We can find two variants of Python 3.5.1 for windows. One is 32-bit, and another is a 64-bit version. And over the past few years, almost all the PCs sold are of 64-bit systems. And here we focus on primarily in 64- bit versions. Also, there are several ways to install the NumPy, and here we are focusing the installation through pip. You can also use the conda package to install it.

Till this date, we cannot find the official NumPy 64 bits available for the different versions of the window and the fact is that NumPy does not officially support 64-bit windows. If you want the NumPy to support 64-bit Windows, then the anticipation of various kind of issues and extensive testing should be done on your own. And one of the possible reason for this is that for creating proper 64-bit builds of NumPy on your windows, there are no adequate open source toolchains which are reliable, and feature-rich. However, the working solution for such problems might be the non-official builds created and provided by Christoph Gohlke in coordination with Intel’s commercial compiler suite. It all depends on your choice that you want to use it or can use these builds. It is always a good job to understand the license impacts at first.

NumPy, An introduction, and brief history

NumPy is an open source software package that is mostly used for the scientific computation of Python. It is because of this reason that it is called NumPy. The Py is for python. Widely known as a library for the Python Programming Language, this package supports large dimensional arrays and matrices including an extensive collection of high-level mathematics functions to operate these arrays and matrices. Along with the scientific uses, NumPy is quite famous for its ability to store the generic data efficiently. It also allows it’s user to define the arbitrary data-types, i.e., derived datatypes such as masked arrays and matrices. Due to this feature, NumPy can integrate seamlessly and speedily with a wide variety of databases. Besides the other features of this package includes a powerful N-dimensional array object, sophisticated functions, shape manipulation, tools for integrating C/C++ and Fortran code, using linear algebra, discrete Fourier transform, and random number simulation capabilities.

 

Actually in the initial stage of the development Python programming language was not used for numerical computing. But it grabbed the attention of the scientific and the engineering community. As the scientific and engineers were very desperate to use it, sadly the array computing package was not defined until 1995. Jim Hugunin at first defined the array computing package with the contributions of other fellow developers. Later in 2005, it was Travis Oliphant who incorporated features of Numarray into Numeric and created NumPy 1.0 (a single array package) with extensive modifications which were released in 2006. The version NumPy 1.5.0 added the support for python 3.

How to install NumPy:

To obtain the correct 64-bit version of Python, you may browse to
http://www.cs.wpi.edu/~cs1004/a16/Resources   

 and download the resulting file to convenient folder or directory. After you downloaded the file right, click on the file python-3.5.1-amd64.exe and selected Run as Administrator to start the installation. Then you can follow the instruction shown on your screen to finish installation or just run the setup.

When you finished installing the Python in your windows, then it is time to install NumPy in your windows. You are suggested to install NumPy package immediately after you install Python 3.5.1.

 

You can download the NumPy package browsing the above-mentioned link;

NumPy-1.11.0-cp35-none-win_amd64.whl

 

For installing it, you can follow the following procedures

Open a Windows Command Prompt and run the command prompt in administrator mode. If you can run command prompt in Administrator mode, then you can see a window resembling figure, and you can identify quickly by looking header of this window which labels as an Administrator Command Prompt.

 

To make the administrator mode if you are using windows seven click on the start button, select All Programs; Accessories; Right-click on “Command Prompt” and select Run as      Administrator.

 

In case of Windows 8 search the command prompt; drag an icon to the desktop and select Run as Administrator.

 

In Windows 10, click windows icon, select All apps; windows system; command prompt; select run as administrator.

 

Open the file where you downloaded

              NumPy-1.11.0-cp35-none-win_amd64.whl

 

 Now you should copy Address of the folder by right-clicking on the address bar and selecting Copy Address as text from the pop-up menu.

 

After that in command prompt windows.

 

Type command “cd” along with space and paste the address which you copied as address .i.e. cd <directory address>. Doing this enables the command prompt point to the address where you downloaded the files.

 

Now you should upgrade pip (Python Installation Program). You can do this by running the command pip install –upgrade pip. You can verify it as pip –version so that it verifies that the pip is up to date and belongs to target CPython Version.

Next, install NumPy. For that type the command

Pip install NumPy-1.11.0-cp35-none-win_amd64.whl

or you can install the name of the downloaded file. Doing this you get a reply:

” Installing collected packages:  NumPy

Successfully installed NumPy-1.11.0 “

Great!. Your job is done. Also, you can find very nice third-party Python Distribution which includes the well tested and commercially supported NumPy builds for 64-bit Windows platforms. We hope this article helped.