1. Univariate

In [38]:
import matplotlib
%matplotlib inline
import seaborn as sns

Let's see how data is distributed as per 'hours-per-week'

In [39]:
adult_data['hours-per-week'].plot.hist(bins=5)
Out[39]:
<matplotlib.axes._subplots.AxesSubplot at 0x183278c6cf8>

Let's check the same for age, but with seaborn library

In [40]:
sns.distplot(adult_data['age'], kde=True)
Out[40]:
<matplotlib.axes._subplots.AxesSubplot at 0x18327971da0>

Let's see the same for 'marital-status' which is categorical variable

In [41]:
adult_data['marital-status'].value_counts().plot( kind='bar')
Out[41]:
<matplotlib.axes._subplots.AxesSubplot at 0x1832793f748>

2. Bivariate

Contingency Table - Let's check how data is ditributed among 'workclass' and 'marital-status' columns

In [42]:
pd.crosstab(index= adult_data['workclass'], columns= adult_data['marital-status'])
Out[42]:
marital-status Divorced Married-AF-spouse Married-civ-spouse Married-spouse-absent Never-married Separated Widowed
workclass
Federal-gov 5 0 9 0 7 0 0
Local-gov 11 0 29 0 19 5 4
Private 96 1 274 10 270 14 21
Self-emp-inc 4 0 25 0 2 1 0
Self-emp-not-inc 9 0 53 1 11 5 1
State-gov 3 0 21 1 8 1 1

Scatter Plot

In [43]:
adult_data.plot.scatter(x='age', y='hours-per-week')
Out[43]:
<matplotlib.axes._subplots.AxesSubplot at 0x18327a85320>

It does give us information here that these 2 variables are not related with each other but sometimes when they are related it will give us the info if they are proporatinal or inversaly proportional.

And then again chi square(categorical variables) or correlation(numeric variables) can be used to find relationship between two variables