Data type detection: integer column with small distinct values as categorical

[jinglinpeng]

Is your feature request related to a problem? Please describe. Currently eda.plot detects a column type based on its pandas dataframe type. Sometimes this may not be ideal. For example, in a dataset the gender column may contain two values: 0 for male and 1 for female. And this will be detected as numerical column, while categorical column makes more sense.

Describe the solution you'd like As a start point, we could handle some simple cases. For example, when a column's dataframe type is integer and its distinct values are smaller than a threshold (we can use the default displayed bars as the threshold), then we detect it as categorical column.

[brandonlockhart]

related to #99

[dovahcrow]

• [ ] Add code to detect ordinal types indtypes.py
• [ ] Handle ordinal plotting through the EDA module.