Not just Python, but the scientific stack as well: https://www.continuum.io/downloads Anaconda is better than other Python Distributions because of its package manager conda.* * Python s own package manager pip requires compilation.
The differences between 2 and 3 are small. Python 3 has many exciting new features. Python 2 is maintained for legacy applications only. No new features are backported. Some libraries have not been updated for seven years, and thus do not support Python 3.
New projects should use Python*. * The most current version of Python But there is legacy Python that still supports your seven year old application if you need it.
PyCharm Community Edition (free) Full IDE: Refactoring Projects Navigation Debugging
Spyder Comes with Anaconda Open Source Matlab-Like: Interactive Console Data Views Debugging
Atom & IPython Text Editor and Command Line Choose your own text editor Choose IPython as a great command line
pip install <package> installs packages.
pip install <package> installs packages. But there is a problem If the package includes C code, pip needs a compiler. And compilers are a BIG BAG OF PAIN.
pip install <package> conda install <package> installs packages. binaries. pythons. languages. but only if it s one of these*: abstract-rendering affine alabaster ansi2html appscript argcomplete astroid astropy azure babel basemap bcolz beautiful-soup binstar binstar-build biopython bitarray blaze blaze blist blockspring blz bokeh boost boto bottleneck bsdiff4 btrees certifi cffi chameleon chest chrpath click cligj cloudpickle clyent cmake colorama conda conda-api conda-build con configobj coverage cryptography cssselect csvkit cubes curl cvxopt cymem cython cytoolz dask datashape datrie dbf decorator dill django docopt docutils dynd-python ecdsa ephe execnet fastcache feedparser fiona flake8 flask flask-login flask-wtf fontconfig freeglut freetype future futures gdal gensim geos greenlet gunicorn h5py hdf5 heapdict holoviews h icu idna iopro ipython itsdangerous jdcal jedi jinja2 joblib jpeg jsonschema lancet-ioam launcher ldap3 libconda libdynd libffi libgdal libnetcdf libpng libsodium libtiff libxml2 libxslt line_profiler llvmlite locket lockfile logilab-common lxml markdown markdown2 markupsafe mathjax matplotlib mccabe mdp meld3 menuinst mingw mistune mock mpmath msg python multimethods multipledispatch murmurhash mysql-connector-python nano natsort ncurses netcdf4 networkx nltk node-webkit nose numba numexpr numpy numpydoc o openpyxl openssl pandas param paramiko partd passlib pastedeploy patchelf patsy pep8 persistent pexpect pillow pip plac ply preshed psutil psycopg2 ptyprocess py pyasn1 pyco pycparser pycrypto pycurl pyflakes pygments pylint pymc pymongo pymysql pyodbc pyopengl pyopengl-accelerate pyopenssl pyparsing pyqt pyramid pyreadline pyserial pysnmp pytables pytest pytest-cache pytest-pep8 python python-dateutil pytz pywget pywin32 pyyaml pyzmq qt quandl queuelib rasterio readline redis redis-py reportlab requests rope sas7bdat scikit-bio scikit-image scikit-learn scipy seaborn semantic_version setuptools sh shapely sip six snowballstemmer snuggs sockjs-tornado spacy sphinx sphinx_rtd_theme s sqlalchemy sqlite sqlparse statsmodels stripe sympy terminado theano thinc tk toolz tornado translationstring twisted ujson unidecode unixodbc unxutils util-linux venusian virtua * it s not that big a deal in practice
conda install <package> installs most packages. pip install <package> installs all other packages. On Windows, if you re desperate, there s also www.lfd.uci.edu/~gohlke/pythonlibs/
For Python: https://www.python.org/ For Numpy/Scipy: https://scipy.org/ For Matplotlib: http://matplotlib.org/ you get the idea. Every package has its own website.
The Python Tutorial https://docs.python.org/3/tutorial/ Dive Into Python 3 http://www.diveintopython3.net/ A Byte of Python (Python 2*) http://www.swaroopch.com/notes/python/ * the differences between 2 and 3 are small