At this month’s meetup Oran Looney gave an hour long presentation about IPython, SciPy, and Pandas. After the meeting, One of the member’s new to the group, Bill Blondeau, sent me an email with feedback. With his permission, I’m posting a slightly edited group of excerpts from their email
TL;DR The Python scientific libraries are very powerful and easy to use. I expect we will be using this stuff more and more as we go forward. Python appears to be gaining ground as the go-to scientific computing solution.
The presenter, Oran Looney, is a Data Scientist with an academic background in physics modeling and visualization. He knows the problem domain professionally. The talk included wry allusions to the pain of dealing with weird heterogeneous data indicating he has first hand practical experience.
Overall it was a good talk for an audience that understands Python but is not familiar with the scientific side. The technologies are very mature and the scientific Python ecosystem is vigorous, as lucidly described in Tal Yakoni’s article.
The talk highlighted the following items:
-
The IPython interactive computing environment, which is an excellent fit for cut-and-try scientific work
-
The SciPy library of open source scientific tools
-
pylab, a user-friendly API package providing clean and straightforward access to the SciPy goodness
-
NumPy, the high-performance multidimensional array package with hefty convenience support methods (linear algebra etc)
-
Pandas, the Python Data Analysis Library
- Specifically Pandas’ DataFrame API
-
MatPlotLib, a powerful 2D plotting and visualization library
-
Anaconda, a scientific-computing-cetnric package manager that, unlike pip, is able to manage non-python packages too.
Thanks to Oran for a great presentation and Bill for this great set of notes!
