Lesson 3. About data types in Python - Data Science for scientists 101

Learning Objectives

At the end of this activity, you will be able to:

  • Understand the structure of and be able to create a vector object in Python.

What You Need

You need Python 3.x and Jupyter notebooks to complete this tutorial. Also you should have an earth-analytics directory setup on your computer with a /data directory with it.

Download Colorado Flood Teaching Data Subset data

To begin working with arrays in python you will first load the numpy library.

import numpy as np

Numpy arrays and data types

An array is a common data structure used in python. An array is defined as a group of values, which most often are either numbers or characters. You can assign this list of values to an object or variable, just like you can for a single value. For example you can create a vector of animal weights:

weight_g = np.array([50, 60, 65, 82])
array([50, 60, 65, 82])

An array can also contain characters:

animals = np.array(['mouse', 'rat', 'dog'])
array(['mouse', 'rat', 'dog'], 

There are many functions that allow you to inspect the content and structure of an array. For instance, len() (short for length) tells you how many elements are in a particular vector:


Array data types

An important feature of an array, is that all of the elements are the same data type. The attribute .dtype shows us the the data type stored within an array:


The function type() shows us the structure of the object.

# View the python object type

You can add elements to an array using the .hstack() function. Below, you add the value 90 to the end of the weight_g object.

# add the number 90 to the end of the vector
weight_g = np.hstack([weight_g, 90])
# add the number 30 to the beginning of the vector
weight_g = np.hstack([30, weight_g])
array([30, 50, 60, 65, 82, 90])

In the examples above, you saw 2 standard data types that Python uses:

  1. "String" and
  2. "Integer".

These are four primative data types that all python objects are built from. They are

  • "Boolean" for TRUE and FALSE
  • "Integer" : whole numbers from negative infinity to infinity
  • "Float" : Rational Numbers ending with decinmal points
  • "String" : Collections of words or characters in single or double quotes

Data type vs. data structure

Data structures in python include arrays, lists, tuples, sets, and dictonaries.

  • "Arrays": In this class you will work with arrays through the numpy modeule, it is important to know that arrays and lists are similar but arrays must be made up of elments that are the same data type.
  • "lists": denoted by [] and contain a series of values, seperated by , they are mutable in python which means they can be changed
  • "Tuple": defined by () and contain a series of values, but tuples are immutable
  • "Sets": a collect of unique objects. Sets are mutable
  • "Dictonaries" list of key value pairs that are denoted by {} a key and its value(s) are seperated by a : and key value pairs are sepearted by commas.

Optional challenge activity

  • Question: What happens when you create a list that contains both numbers and character values? Give it a try and write down the answer.
  • Question: What will happen in each of these examples? (hint: use type() to check the data type of your objects):
num_char = np.array([1, 2, 3, 'a'])

num_logical = np.array([1, 2, 3, True])

char_logical = np.array(['a', 'b', 'c', True])

tricky = np.array([1, 2, 3, '4'])
  • Question: Why do you think it happens?
  • Question: Can you draw a diagram that represents the hierarchy of the data types?
animals = np.array(["mouse", "rat", "dog", "cat"])


Subset arrays

If you want to extract one or several values from a vector, you must provide one or several indices in square brackets. For instance:

animals[[2, 3]]
array(['dog', 'cat'], 

0 vs 1-based Indexing

Python and other languages such as C++, Java, and Perl count from 0. The first element of a list in python will be accessed with listname[0]. Programming languages like Fortran, MATLAB, and R start counting at 1. So, in comparison, in R, you would use listname[1].

Subsetting using a series of sequential numbers in Python, is also different from a language like R. In python, you the end n umber in the subset is 1 index BEYOND the last index value. So, if you want the first three elements of a list you would run listname[0:3] and that would return the elements at positions 0, 1, and 2.

You can subset arrays too. For instance, if you want to select only the values that are greater than 50:

weight_g > 50
array([False, False,  True,  True,  True,  True], dtype=bool)

Notice that the command above returns a BOOLEAN (TRUE / FALSE) array. You can then use that array to select all objects in your weight_g array that are greater than 50 as follows:

# select only the values greater than 50
weight_g[weight_g > 50]
array([60, 65, 82])

You can combine multiple tests using & (both conditions are TRUE, AND) or | (at least one of the conditions is TRUE, OR):

# select objects that are EITHER less than 30 OR greater than 50
weight_g[(weight_g < 30) | (weight_g > 50)]
array([60, 65, 82])
# select objects that are greater than or equal to 30 OR equal to 21
weight_g[(weight_g >= 30) & (weight_g == 21)]
array([], dtype=int64)

Notice that you use two == signs to designate equal to so as not to confuse equals to with the assignment operator which is also = in Python.

animals = np.array(['mouse', 'rat', 'dog', 'cat'])
animals[(animals == 'cat') | (animals == 'rat')]
array(['rat', 'cat'],

The function intersection() allows you to test if a value is found in an array of values:

# select objects in the animals array that are within the array [`rat`, `cat`, `dog`, `duck`]
set(animals).intersection(set(['rat', 'cat', 'dog', 'duck']))
{'cat', 'dog', 'rat'}

Optional challenge

  • Can you figure out why "four" > "five" returns TRUE?


Leave a Comment