Making Data Management Decisions

assignment.py

#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Thu Jan 27 11:30:33 2017

@author: mllearner
"""

import pandas
import numpy

data = pandas.read_csv('gapminder.csv')
print len(data)         # Number of observations
print len(data.columns) # Number of variables

# IMPORTANT NOTE: Since I know all my variables are lower case, I skipped the part where
# we manually change every char upper case.
pandas.set_option('display.float_format', lambda x: '%f' %x)

datavariable1 = "femaleemployrate"
datavariable2 = "internetuserate"
datavariable3 = "urbanrate"

data[datavariable1] = data[datavariable1].convert_objects(convert_numeric=True)
data[datavariable2] = data[datavariable2].convert_objects(convert_numeric=True)
data[datavariable3] = data[datavariable3].convert_objects(convert_numeric=True)

# Subset of observations with no empty data. You can aswell add more complicated logic
# statements to narrow our results down.
sub1 = data[(data['femaleemployrate'] > 0) & (data['internetuserate'] > 0)]

sub2 = sub1.copy()

print "counts for femaleemployrate // switched to polityscore due to readibilty"
c11 = sub2['polityscore'].value_counts(sort=False)
print c11

sub2['polityscore'] = sub2['polityscore'].replace('-2', numpy.nan)
ctest = sub2['polityscore'].value_counts(sort=False, dropna=False)
print ctest
# we can clearly see that if we use femaleemployrate instead of polityscore
# we'll end up seeing no NAN values, since floating point issue appears.
# I instead switched to polityscore.

# I'm gonna check for null variables, since my dataset includes them.
cnull = sub2['polityscore'].head(25).isnull()
# none of them is null. oh polityscore has 2, 4 and 210. bingo.
sub2['urbanrate'] = sub2['urbanrate'] * sub2['urbanrate']
# ok we successfully multiplied.