Using proper data types

Analyzing Police Activity with pandas

Kevin Markham

Founder, Data School

Examining the data types

ri.dtypes

stop_date             object
stop_time             object
driver_gender         object
...                      ...
stop_duration         object
drugs_related_stop      bool
district              object

object: Python strings (or other Python objects)
bool: True and False values
Other types: int, float, datetime, category

Why do data types matter?

Affects which operations you can perform
Avoid storing data as strings (when possible)
- int, float: enables mathematical operations
- datetime: enables date-based attributes and methods
- category: uses less memory and runs faster
- bool: enables logical and mathematical operations

Fixing a data type

apple

      date   time   price
0  2/13/18  16:00  164.34
1  2/14/18  16:00  167.37
2  2/15/18  16:00  172.99

apple.price.dtype

dtype('O')

apple['price'] = 
  apple.price.astype('float')

apple.price.dtype

dtype('float64')

Dot notation: apple.price
Bracket notation: apple['price']

Let's practice!

Analyzing Police Activity with pandas