Will it Python? ARM Chapter 5: Logistic models of well-switching in Bangladesh

UPDATE 1/15/2014: This blog is no longer in service.

This post is now located at: http://slendermeans.org/arm-ch5.html

Thanks,
-c.

This entry was posted in Uncategorized. Bookmark the permalink.

4 Responses to Will it Python? ARM Chapter 5: Logistic models of well-switching in Bangladesh

  1. Thomas Haslwaner says:

    Very nice work!
    There are two bugs, though (at least if you use statsmodels 0.5.0)
    1) In your calls to “logit”, you have to replace “df=df” with “data=df”
    2) To get the marginal effects, you have to use
    “model2.get_margeff(at = ‘mean’).summary()”
    instead of “model2.margeff(at=’mean’)

    Also, make sure to use pandas 10.1 – otherwise you will have problems with the “read_cvs”

    • Carl says:

      Thomas —
      Everything works fine for me under statsmodels 0.5.0 (0.5.0.dev-1bbd4ca to be exact). Not sure why the notebook as-is is buggy for you.

      And what problem are you finding with read_csv under pre 0.10.1 pandas? I don’t have a problem with 0.10 (and I think this may have even worked under 0.9.x).

  2. Thomas Haslwaner says:

    oh, and another thing: the interaction distance:arsenic is not significant (P=0.08)

Comments are closed.