[Chicago] Fun in learning ML

Ryan Nelson rnelsonchem at gmail.com
Wed Feb 11 03:37:03 CET 2015 

Sorry to be a little late to the party here. Just now looking at this.

For scientific Python packages (Numpy, Scipy, Pandas, Statsmodels, etc.),
I've had very good luck using "conda" to create test environments and to
install all of a packages dependencies. These environments are like
virtualenvs, and they can be installed without admin privileges in a
directory of your choice without affecting system-installed libraries. The
Anaconda installer contains all the packages (probably more than you need),
or use Miniconda to get a minimal environment. Using conda, you could
create a Statsmodels/Pandas environment (called "stats") with all the
dependencies as such:

$ conda create -n stats python=3 statsmodels pandas
$ source activate stats
stats>$ # Do stuff
stats>$ source deactivate

I don't use PyCharm, but with conda, you can install Spyder. Otherwise,
there is documentation for using other IDEs including PyCharm:
http://docs.continuum.io/anaconda/ide_integration.html

Anaconda Download: http://continuum.io/downloads
Miniconda Download: http://conda.pydata.org/miniconda.html
Conda Docs: http://conda.pydata.org/docs/index.html

Ryan


> Ahhh.. this is someone elses code you are trying to run, and you don't have
> the dependencies installed.
> It may make you feel good to know this is a problem that spans decades and
> pretty much every language.    Now that you feel better, that doesn't
> really solve your problem.
> my guess is
> 1. statsmodels is expecting an older version of pandas than what you
> installed.
> 2. someone stepped on some name space.
> import foo
> foo=1
> foo.bar()
> when those 3 lines are not next to each other, it is a real pain to figure
> out what happened.
> Lets hope it is #1.
>
>
>
>
>
>
> On Mon, Feb 2, 2015 at 5:07 PM, Randy Baxley <randy7771026 at gmail.com>
> wrote:
> > That statement is buried in a library called in statsmodels / pandas by
> > this piece of code that should work at least down to where reg(x) is
> called
> > because it was working for the guy who gave it to us in a breakout at
> > chihacknight:
> >
> > import statsmodels.api as sm
> > # import statsmodels.api, numpy, scipy, matplotlib, pandas, sympy, nose
> as sm
> >
> > y = [5, 6, 5, 7, 7, 8, 9, 9, 11, 12, 15]
> >
> > print 'y', y
> >
> > x = [1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 4]
> >
> > print 'x', x
> >
> > x = sm.add_constant(x)
> >
> > print x
> >
> > model = sm.api.OLS(y,x) # does an ordinary least squared regression
> > results = model.fit()
> >
> > p = results.params
> >
> > print p
> >
> > def reg(x):
> >
> >     return p[0] + p[1]*x
> >
> > print reg(3)
> >
> >
> >
> > On Mon, Feb 2, 2015 at 2:19 PM, Carl Karsten <carl at personnelware.com>
> > wrote:
> >
> >> the link I gave might be close, but I doubt it is an exact hit.
> >>
> >> Where/why do you have this line: (why do you think it should work?)
> >>
> >> from pandas.core.api import get_dummies
> >>
> >> On Mon, Feb 2, 2015 at 1:46 PM, Randy Baxley <randy7771026 at gmail.com
> >
> >> wrote:
> >>
> >>> Yep and my mind has a lot of fun with the idea of get_dummies.
> >>>
> >>> I cut off the traceback one line too soon.  Maybe the link Carl gave
> >>> will have an answer but just for completeness here is the traceback
> with
> >>> that line included.
> >>>
> >>> >>> import statsmodels.api as sm
> >>> Traceback (most recent call last):
> >>>   File "<stdin>", line 1, in <module>
> >>>   File "/usr/local/lib/python2.7/dist-packages/statsmodels/api.py",
> line
> >>> 13, in <module>
> >>>     from .discrete.discrete_model import (Poisson, Logit, Probit,
> >>>   File
> >>>
> "/usr/local/lib/python2.7/dist-packages/statsmodels/discrete/discrete_model.py",
> >>> line 40, in <module>
> >>>     from pandas.core.api import get_dummies
> >>> ImportError: cannot import name get_dummies
> >>>
> >>>
> >>> On Mon, Feb 2, 2015 at 1:02 PM, Carl Karsten <carl at
> personnelware.com>
> >>> wrote:
> >>>
> >>>> You didn't post the error, but earlier I thnk I see it:
> >>>>
> >>>>     from pandas.core.api import get_dummies
> >>>> ImportError: cannot import name get_dummies
> >>>>
> >>>> Assuming that is still the error..
> >>>>
> >>>> That says there is no get_dummies in pandas.core.api.
> >>>>
> >>>> I did find this:
> >>>>
> >>>>
> >>>>
> http://pandas.pydata.org/pandas-docs/stable/generated/pandas.core.reshape.get_dummies.html
> >>>>
> >>>> So something has get_dummies.  but given the description I expect many
> >>>> things have get_dummies.
> >>>>
> >>>> I'll let you find the docs that relate to what you are trying to do.
> >>>>
> >>>>
> >>>>
> >>>> On Mon, Feb 2, 2015 at 12:35 PM, Randy Baxley <randy7771026 at
> gmail.com>
> >>>> wrote:
> >>>>
> >>>>> True that.
> >>>>>
> >>>>> The error comes up the same in terminal using python.
> >>>>>
> >>>>> >>> import statsmodels.api as sm
> >>>>> Traceback (most recent call last):
> >>>>>   File "<stdin>", line 1, in <module>
> >>>>>   File "/usr/local/lib/python2.7/dist-packages/statsmodels/api.py",
> >>>>> line 13, in <module>
> >>>>>     from .discrete.discrete_model import (Poisson, Logit, Probit,
> >>>>>   File
> >>>>>
> "/usr/local/lib/python2.7/dist-packages/statsmodels/discrete/discrete_model.py",
> >>>>> line 40, in <module>
> >>>>>     from pandas.core.api import get_dummies
> >>>>>
> >>>>>
> >>>>> On Mon, Feb 2, 2015 at 12:12 PM, Carl Karsten <carl at
> personnelware.com>
> >>>>> wrote:
> >>>>>
> >>>>>> good.  pandas is installed and available.
> >>>>>>
> >>>>>> Now try the line or lines that are giving you trouble.  Or if you
> >>>>>> need to run your program to get an error, do that.
> >>>>>>
> >>>>>> The idea is very simple text based cause and effect.
> >>>>>>
> >>>>>> An IDE (PyCharm) will help you write code, but does not help us
> debug
> >>>>>> your code.  It just makes things harder for us, so when asking for
> help
> >>>>>> with your code it is best to remove the IDE component.
> >>>>>>
> >>>>>> Unless you need help using the IDE, but that's a different problem
> >>>>>> than how do I use pandas.
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> On Mon, Feb 2, 2015 at 11:56 AM, Randy Baxley <randy7771026 at
> gmail.com
> >>>>>> > wrote:
> >>>>>>
> >>>>>>> user at freegeekchicago:~$ python
> >>>>>>> Python 2.7.3 (default, Dec 18 2014, 19:10:20)
> >>>>>>> [GCC 4.6.3] on linux2
> >>>>>>> Type "help", "copyright", "credits" or "license" for more
> >>>>>>> information.
> >>>>>>> >>> import pandas
> >>>>>>> >>>
> >>>>>>>
> >>>>>>>
> >>>>>>> On Mon, Feb 2, 2015 at 10:29 AM, Carl Karsten <
> >>>>>>> carl at personnelware.com> wrote:
> >>>>>>>
> >>>>>>>> IDEs and screen shots make it harder to converse over email.
> >>>>>>>>
> >>>>>>>> Use the simple text based python repl so you can cut/paste text.
> >>>>>>>> I don't have it installed, so this is what I get:
> >>>>>>>>
> >>>>>>>> carl at twist:~$ python
> >>>>>>>> Python 2.7.6 (default, Mar 22 2014, 22:59:56)
> >>>>>>>> [GCC 4.8.2] on linux2
> >>>>>>>> Type "help", "copyright", "credits" or "license" for more
> >>>>>>>> information.
> >>>>>>>>
> >>>>>>>> >>> import pandas
> >>>>>>>> Traceback (most recent call last):
> >>>>>>>>   File "<stdin>", line 1, in <module>
> >>>>>>>> ImportError: No module named pandas
> >>>>>>>> >>>
> >>>>>>>>
> >>>>>>>> Do that and post the results.
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> On Sun, Feb 1, 2015 at 1:18 AM, Randy Baxley <
> >>>>>>>> randy7771026 at gmail.com> wrote:
> >>>>>>>>
> >>>>>>>>> It was worth a try so I did it.  I think Pandas is installed but
> >>>>>>>>> on the path highlighted and in yellow at the top in fir.
> >>>>>>>>> st attachment
> >>>>>>>>> Second attachment shows where I think Python is looking for
> Pandas,
> >>>>>>>>>
> >>>>>>>>> On Sat, Jan 31, 2015 at 10:23 PM, Carl Karsten <
> >>>>>>>>> carl at personnelware.com> wrote:
> >>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> On Sat, Jan 31, 2015 at 10:02 PM, Randy Baxley <
> >>>>>>>>>> randy7771026 at gmail.com> wrote:
> >>>>>>>>>>
> >>>>>>>>>>>     from pandas.core.api import get_dummies
> >>>>>>>>>>> ImportError: cannot import name get_dummies
> >>>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> pandas isn't installed right. (maybe not at all)
> >>>>>>>>>>
> >>>>>>>>>> my guess:
> >>>>>>>>>> sudo easy_install pandas
> >>>>>>>>>>
> >>>>>>>>>> --
> >>>>>>>>>> Carl K
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>> _______________________________________________
> >>>>>>>>>> Chicago mailing list
> >>>>>>>>>> Chicago at python.org
> >>>>>>>>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>>>>>>>
> >>>>>>>>>>
> >>>>>>>>>
> >>>>>>>>> _______________________________________________
> >>>>>>>>> Chicago mailing list
> >>>>>>>>> Chicago at python.org
> >>>>>>>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> --
> >>>>>>>> Carl K
> >>>>>>>>
> >>>>>>>>
> >>>>>>>> _______________________________________________
> >>>>>>>> Chicago mailing list
> >>>>>>>> Chicago at python.org
> >>>>>>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>>>>>
> >>>>>>>>
> >>>>>>>
> >>>>>>> _______________________________________________
> >>>>>>> Chicago mailing list
> >>>>>>> Chicago at python.org
> >>>>>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>>>>
> >>>>>>>
> >>>>>>
> >>>>>>
> >>>>>> --
> >>>>>> Carl K
> >>>>>>
> >>>>>>
> >>>>>> _______________________________________________
> >>>>>> Chicago mailing list
> >>>>>> Chicago at python.org
> >>>>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>>>
> >>>>>>
> >>>>>
> >>>>> _______________________________________________
> >>>>> Chicago mailing list
> >>>>> Chicago at python.org
> >>>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>>
> >>>>>
> >>>>
> >>>>
> >>>> --
> >>>> Carl K
> >>>>
> >>>>
> >>>> _______________________________________________
> >>>> Chicago mailing list
> >>>> Chicago at python.org
> >>>> https://mail.python.org/mailman/listinfo/chicago
> >>>>
> >>>>
> >>>
> >>> _______________________________________________
> >>> Chicago mailing list
> >>> Chicago at python.org
> >>> https://mail.python.org/mailman/listinfo/chicago
> >>>
> >>>
> >>
> >>
> >> --
> >> Carl K
> >>
> >>
> >> _______________________________________________
> >> Chicago mailing list
> >> Chicago at python.org
> >> https://mail.python.org/mailman/listinfo/chicago
> >>
> >>
> >
> > _______________________________________________
> > Chicago mailing list
> > Chicago at python.org
> > https://mail.python.org/mailman/listinfo/chicago
> >
> >
>
> --
> Carl K
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/chicago/attachments/20150210/737d46ca/attachment.html>

More information about the Chicago mailing list