The integration, verification, and sometimes certification of complex systems involves the development and execution of test cases. These test cases generate data that is analyzed to determine if the system meets its requirements. Let’s look at the eight reasons why we prefer Python as our data analysis language of choice.
1. It is open source
It’s no secret that the team at Applied Dynamics has love for open source technology. We’ve been around long enough to get burned by the selection of proprietary technology that eventually went the way of the dodo bird. Our software, our hardware designs, and our development processes were built up around technology, some of which eventually died or became too costly compared to better alternatives. Our only choice was to extract that software technology from our world and replace it with something better, almost always open source. The zero dollar licensing cost of getting started is also a pretty nice feature of Python as an open source platform for data analysis.
2. It is object oriented
The patterns within a test and data analysis language beg for an object oriented implementation. A test is comprised of a collection of test items. There is a common set of properties and attributes that can be identified across all test items, lending itself to inheritance. Analysis methods are built up as a collection of method items whose structure is dependent on the dataset schema. Then there is the evidence generation portion of the test and analysis effort, another set of data structures that naturally lend themselves to a well-defined data class architecture. Building up data classes in Python is intuitive and powerful. It’s the right thing to do; and it’s the right way to do it
3. and 4. NumPy and SciPy, not necessarily in that order
Linear and matrix algebra, integration, optimization, interpolation, Fast Fourier transforms, signal processing, sparse eigenvalues, statistics, and control systems analysis are the tip of the two icebergs that are SciPy and NumPy. One of the beautiful underlying philosophical tenants of the open source movement is the idea that the value of a given piece of software falls over time and eventually becomes free and public domain. SciPy and NumPy are two dramatic examples of tremendous value that is now free for the benefit of all.
5. Finding talent
A key consideration for the selection of any test and analysis language is access to experienced practitioners. Experienced Python users are easy to find. Its heavy application in web-based software is probably one reason why there are so many Python coders out there. We’ve also noticed a fast emergence of Python education within US universities and many engineering and computer science grads have hands-on experience right out of school.
The ability to extend your test and analysis through the use of platform-neutral API’s is an aspect of Python limited only by your imagination. We like to think of Python as a sandbox allowing technology to be brought together however the heck you wanna bring it!
7. Embed – Ability
Once upon a time Microsoft offered and encouraged the embedding of an interpreted VBA scripting module for desktop applications written in MFC or .NET. That piece of technology soon disappeared and Visual Basic immediately became a less appropriate embedded scripting language. By this time, Python was already positioning itself as a better embedded scripting approach. And wow, has the value proposition grown. The list of popular, commercial software applications who’ve embedded Python enabling their functionality to be scripted, is longer than this top 8 list article. LibreOffice is transitioning their default from Java to Python. The Raspberry Pi has adopted Python as the principal user programming language.
8. Easy to learn
Python is easy to learn. Anyone with experience programming in C or Java will find it as easy, and more rewarding, than falling down stairs. The internet offers an endless supply of educational resources to learn Python including free Python training at Codecademy.com.