Commit 57c7547

**********************

Intro to Numpy / Scipy

**********************

- Chris Barker

(April 29, 2014)

Scipy

-----

The scipy "Stack" is a collection of core packages used for scientific / numerical computing.

http://www.scipy.org/stackspec.html

Many other domain-specific packages area available:

Core "stack" is what most people will want, regardless of domain.

What's in the scipy stack?

--------------------------

* Python (http://www.python.org)

* NumPy (http://www.numpy.org)

* SciPy library (http://www.scipy.org)

* Matplotlib (http://matplotlib.org/)

* IPython (http://ipython.org/)

* nose (https://nose.readthedocs.org)

* pandas (http://pandas.pydata.org/)

* Sympy (http://sympy.org/)

numpy is the core package that the rest of the scipy stack is built on. Mostly what I'll talk about here.

http://www.numpy.org/

An N-Dimensional array object

demo: ``mem_struct.ipynb``

* Bool -- 8 bit

demo: ``object.ipnb``

Compound dtypes

Still Homogeneous: Array of structs.

* ``ones(), zeros(), empty(), arange(), linspace(), logspace()``

( Default dtype: ``np.float64`` )

* ``array(), asarray()`` ( Build from any sequence )

From binary data:

* ``fromstring(), frombuffer(), fromfile()``

Text File I/O

--------------

Loading from text (CSV, etc):

* ``np.loadtxt``

* ``np.genfromtxt`` ( a few more features )

Saving as text (CSV):

* ``np.savetxt()``

Numpy Persistence:

``np.tofile() / np.fromfile()``

``np.savez()`` -- numpy zip format

* NetCDF4

<title>Intro to Numpy / Scipy &mdash; numpy</title>

<article class="slide level-1" id="intro-to-numpy-scipy">

<h1>Intro to Numpy / Scipy</h1>

<ul class="simple">

<li>Chris Barker</li>

<p>(April 29, 2014)</p>

<article class="slide level-2" id="scipy">

<p>The scipy &quot;Stack&quot; is a collection of core packages used for scientific / numerical computing.</p>

<p><a class="reference external" href="http://www.scipy.org/stackspec.html">http://www.scipy.org/stackspec.html</a></p>

<p>Many other domain-specific packages area available:</p>

<div>Core &quot;stack&quot; is what most people will want, regardless of domain.</div></blockquote>

<article class="slide level-2" id="what-s-in-the-scipy-stack">

<h2>What's in the scipy stack?</h2>

<ul class="simple">

<li>Python (<a class="reference external" href="http://www.python.org">http://www.python.org</a>)</li>

<li>NumPy (<a class="reference external" href="http://www.numpy.org">http://www.numpy.org</a>)</li>

<li>SciPy library (<a class="reference external" href="http://www.scipy.org">http://www.scipy.org</a>)</li>

<li>Matplotlib (<a class="reference external" href="http://matplotlib.org/">http://matplotlib.org/</a>)</li>

<li>IPython (<a class="reference external" href="http://ipython.org/">http://ipython.org/</a>)</li>

<li>nose (<a class="reference external" href="https://nose.readthedocs.org">https://nose.readthedocs.org</a>)</li>

<li>pandas (<a class="reference external" href="http://pandas.pydata.org/">http://pandas.pydata.org/</a>)</li>

<li>Sympy (<a class="reference external" href="http://sympy.org/">http://sympy.org/</a>)</li>

<article class="slide level-2" id="numpy">

<h2>numpy</h2>

<p>numpy is the core package that the rest of the scipy stack is built on. Mostly what I'll talk about here.</p>

<p>Not just for lots of numbers!

(but it's great for that!)</p>

<p><a class="reference external" href="http://www.numpy.org/">http://www.numpy.org/</a></p>

<p>An N-Dimensional array object</p>

<p>demo: <tt class="docutils literal"><span class="pre">mem_struct.ipynb</span></tt></p>

<li><p class="first">Bool -- 8 bit</p>

<p>demo: <tt class="docutils literal"><span class="pre">object.ipnb</span></tt></p>

<article class="slide level-2" id="compound-dtypes">

<h2>Compound dtypes</h2>

<dd><p class="first last">Still Homogeneous: Array of structs.</p>

<dl class="docutils">

<dt>From scratch:</dt>

<dd><ul class="first simple">

<li><tt class="docutils literal"><span class="pre">ones(),</span> <span class="pre">zeros(),</span> <span class="pre">empty(),</span> <span class="pre">arange(),</span> <span class="pre">linspace(),</span> <span class="pre">logspace()</span></tt></li>

<p class="last">( Default dtype: <tt class="docutils literal"><span class="pre">np.float64</span></tt> )</p>

<dt>From sequences:</dt>

<dd><ul class="first last simple">

<li><tt class="docutils literal"><span class="pre">array(),</span> <span class="pre">asarray()</span></tt> ( Build from any sequence )</li>

<dt>From binary data:</dt>

<dd><ul class="first last simple">

<li><tt class="docutils literal"><span class="pre">fromstring(),</span> <span class="pre">frombuffer(),</span> <span class="pre">fromfile()</span></tt></li>

<article class="slide level-2" id="text-file-i-o">

<h2>Text File I/O</h2>

<p>Loading from text (CSV, etc):</p>

<div><ul class="simple">

<li><tt class="docutils literal"><span class="pre">np.loadtxt</span></tt></li>

<li><tt class="docutils literal"><span class="pre">np.genfromtxt</span></tt> ( a few more features )</li>

<p>Saving as text (CSV):</p>

<div><ul class="simple">

<li><tt class="docutils literal"><span class="pre">np.savetxt()</span></tt></li>

<article class="slide level-2" id="numpy-persistence">

<h2>Numpy Persistence:</h2>

<p><tt class="docutils literal"><span class="pre">np.tofile()</span> <span class="pre">/</span> <span class="pre">np.fromfile()</span></tt></p>

<p><tt class="docutils literal"><span class="pre">np.savez()</span></tt> -- numpy zip format</p>

<dt>netcdf</dt>

<dd><ul class="first last simple">

<li>NetCDF4</li>