To submit a user story, send to: info@numfocus.org.
Circuitscape Protects Endangered Species Using PyAMG
Scientists map the “current” of mountain lions moving between mountain ranges. Blue shows areas of low current density, which are expected to have low densities of dispersing mountain lions; yellow designates movement bottlenecks, which are most vulnerable to habitat destruction. Destroying high-flow habitat can isolate populations and endanger their survival.
Image: Brad McRae and Brett Dickson
PyAMG is an open-source numerical package of multigrid and Krylov solvers for sparse matrices in Python. One of the biggest users of PyAMG is Circuitscape. This open-source software program borrows algorithms from electronic circuit theory to predict patterns of movement, gene flow, and genetic differentiation among plant and animal populations in heterogeneous landscapes. Circuit theory complements least-cost path approaches because it considers effects of all possible pathways across a landscape simultaneously.
Circuitscape was written by Brad McRae and Viral Shah. Brad is an Ecologist with the The Nature Conservancy in Seattle. He works on habitat connectivity conservation, climate change, and landscape genetics. Viral is a Senior Scientist with Interactive Supercomputing and a Visiting Scientist at the University of California, Santa Barbara. He develops systems for scientific computing for work and for fun.
For additional information on Circuitscape:
The authors presentation at SciPy 2008: http://www.circuitscape.org/Circuitscape/Pubs_files/Shah_McRae_Circuitscape_Python_Scipy08.pdf.
Conservation Magazine: Circuitous Routes: Circuit Theory Guides Wildlife Corridor Design by Eric Wagner
Wired Magazine: Scientist Employs ‘Circuit Theory’ to Protect Endangered Species by Carl Zimmer
Convergence Magazine: Where the Wild Things Are by Anna Davison
Earthtimes.org: Where Mountain Lions Roam: Star-P Helps Decipher Threatened Wildlife Migration
Python in Astronomy
Mark Garlick/University of Warwick Black Hole
On 28 March, NASA’s Swift satellite observed a flash of gamma rays brighter than anything astronomers had seen before. It soon became evident that the event wasn’t a typical gamma ray burst, an emission of high-energy radiation that often accompanies a supernova explosion. The flash didn’t die out but was sustained for weeks, and although it has faded in intensity, it is still going strong 2½ months later. Two papers published online today in Science provide an explanation for this luminous surprise. The flare is in fact a high-energy jet of radiation produced by a star falling into a black hole at the center of a galaxy 4 billion light-years away. The reason the flare is so bright is that the jet is pointed straight in the direction of Earth. And it’s sustained becausethe black hole is consuming the star gradually. “That’s because as the black hole rips the star apart, the mass swirls around like water going down a drain, and this swirling process releases a lot of energy,” says Joshua Bloom, an astronomer at the University of California, Berkeley, and lead author of one of the two papers. Bloom expects the flare to fade out over the next year.
http://news.sciencemag.org/sciencenow/2011/06/scienceshot-powerful-jet-being.html?ref=hp
by Yudhijit Bhattacharjee on 16 June 2011, 2:00 PM
For related story and photo, http://discovermagazine.com/2012/jan-feb/14.
Tokyo SciPy
Chairman of NumFOCUS’ Board of Director’s,Travis Oliphant, spoke at the March meeting of Tokyo SciPy via live webcast. Topics he covered included the history of SciPy, where NumPy should go next, and Numba. Questions were asked regarding sparse matrix integration with NumPy, and packaging.
Tokyo SciPy member, Syoyo, gave Travis the link to his code for simulating smoke, http://code.google.com/p/smoke3d/. Perhaps at some point, Syoyo will be able to make that code fast using Python.
Despite a late night, due to the fourteen hour time difference between Austin and Tokyo, Travis spoke for over an hour. He later stated, “…it was nice to see that SciPy and NumPy have a user community in Japan.”
Thank you for introducing Tokyo.Scipy. We are happy to see a chronicle of NumPy.
Thanks! I had tried to order this but it was out of stock, now arriving in 3 days, hope the Amazon link rattels your tip jar.There is an interesting interview at the Python411 podcast with the author.Also, if people have trouble with installation of dependencies, they might try a program called Reinteract, which takes an interesting approach to python visualization it renders images inline, alongside the interpreter. See reinteract.org.
Joshua Bloom, the Berkeley astronomer quoted in the article, will be one of the keynote speaks at the SciPy conference this summer. It should be a great talk (at a great conference)!
(I think a previous ceommnt along these lines might have been spam-trapped for excess links here is a googleable version)A great print book: Core Python Programming by Wesley Chun author is a gifted teacher as well as programmer.Some great free online resources: How to Think Like a Computer Scientist: Learning with Python By Allen Downey of Olin College really outstanding A Byte of Python The official Python tutorialThe tutorial section of Python411, google that or awaretek python tutorials If you already have some fundamentals of programming you can read Mark Pilgrim’s Dive Into Python . It seems oriented towards people with C++ and Java backgrounds. It is a little heavy on things like xml and regular expressions for an introduction to programming, reflecting its original purpose.Your time spent learning this stuff is insanely well invested, you will be amazed.
Flicked through the table of cotnnets looks great! As an engineer/scientist I made the plunge a couple of years ago and switched from Matlab to python. As a user, it’s been great to see the python scientific stuff develop into a mature and stable product suite. The ONE gripe I have at the moment is 3D plotting 2D is awesome with matplotlib, and mayavi/VTK etc. looks great for fancy 3D stuff, but if I just want to draw up a quick 3D surface from a numpy array, it takes way too much effort.Also, I stumbled across python(x,y) recently makes installing a VTK based visualisation system a heck of a lot easier than it used to be (I tried on windows a while ago, and gave up with unmet package dependencies etc.).
Will you be using an algorithm wtihin Python to make the random selection? +)Also, in the same vein, you may try the project, which makes excellent use of Python to provide a common interface to Maxima, NumPy, SciPy, Matplotlib, as well as Maple, Mathematica, and other commercial ventures. Additionally, the Sage developers have built their own procedures and function groups. All of the open-source pieces noted above are bundled in a single install, with very minimal fuss.From the 4.01 release notes:In version 4.0.1, doctest coverage increased by 0.3%, while an additional 177 functions have been added. The overall weighted coverage score is now at 77.2% and the total number of functions is at 21,988.21,988. Nice.
I’m a little sseuriprd that you had trouble installing the dependencies. Have you tried to install from source? There are binary packages for Linux and Windows that make installing very easy, I’m not sure how it would be on a Mac, though. Besides EPD, suggested above, there is also Python(x,y) (pythonxy.com) that bundles python, scipy, mayavi, etc in one package.I wonder how useful the book would be for someone already familiar with matplotlib and scipy. Scipy documentation is very sketchy in some places, so the book might be valuable even for those.
It was mostly becsuae I’ve been using Apple’s packaged Python, and Scipy binary wanted the system version. So I tried installing from source to get around installing Python, but that sorta sucked. So after fussing with that, I had to get all my paths and dependencies straightened out. But yeah, once I took care of that, I think all of the binary packages took care of things. I don’t remember exactly. It was late