learn-co-curriculum / ds-numpy-lab

Numpy lab

Introduction

Numpy is one of the main libraries for performing scientific computing in Python. Using Numpy, you can create high-performance multi-dimensional arrays, and several tools to work with these arrays.

A numpy array can store a grid of values. All the values must be of the same type. numpy arrays are n-dimensional, and the number of dimensions is denoted the rank of the numpy array. The shape of an array is a tuple of integers which hold the size of the array along each of the dimensions.

For more information on numpy, we refer to www.numpy.org.

Objectives

You will be able to:

• Create a Numpy array
• Understand Subsetting
• Understand Higher Dimensional Numpy Arrays

To work with numpy arrays, you should import numpy first. The naming convention for numpy is to import it as np.

Numpy array creation and basic operations

None

Now, create your first numpy array. Often, one would create a list and then convert it to a numpy array. We will create a first numpy array containing the population of the 5 biggest american cities (in millions). The numbers are 8.6, 3.9, 2.7, 2.1 and 1.6 respectively. Store it in city_pop.

For your information, these cities are NYC, LA, Chicago, Houston and Philadelphia respectively.

city_pop = None
city_pop

Look at the type of city_pop and confirm that it is a list.

None

list

now, convert city_pop to a numpy array and store it in city_pop_np

city_pop_np = None

confirm the data type again

None

numpy.ndarray

Now we want to make the numpy array in a way that it reflects the population to the actual unit. You can easily do this by multiplying your entire array with 1000000.

city_pop_m_np= None
city_pop_m_np

array([8600000., 3900000., 2700000., 2100000., 1600000.])

This was easy! Now let's look at another numpy array which stores the surface area of these five cities (in square miles).

city_size_np= None
city_size_np

array([468, 503, 234, 627, 142])

We're interested in finding the population density for the four cities. This is one of the perks of numpy arrays: you simply divide the two arrays by each other!

city_density_np = None
city_density_np

array([18376.06837607,  7753.47912525, 11538.46153846,  3349.28229665,
       11267.6056338 ])

It is no surprise that NYC is very densely populated where the population density of Houston is much lower.

Numpy array subsetting

Imagine you want to select certain elements from a numpy array. This can easily be done using square brackets. Now, go ahead and select the fourth element in the city_density_np array.

density_houston_np = None
density_houston_np

3349.282296650718

Similarly, you can use colons to select a range of items. Use a colon to select the first three densities.

first_three_np = None
first_three_np 

array([18376.06837607,  7753.47912525, 11538.46153846])

Next, we want to select the densities that are higher than 10000 and store them in the object high_density_np

high = None
high_density_np = None
high_density_np

array([18376.06837607, 11538.46153846, 11267.6056338 ])

great!

Higher dimensional numpy arrays

city_pop_m_np, city_size_np, and city_density_np are one-dimensional numpy arrays. This is pretty straightforward, but can be verified by using the attribute .shape.

None

(5,)

This indicates that we have a 1-dimensional numpy array with 5 elements. Now, let's group our three arrays together (population, then size, then density) and see what happens if we call .shape again.

cities_np = None
cities_np

array([[8.60000000e+06, 3.90000000e+06, 2.70000000e+06, 2.10000000e+06,
        1.60000000e+06],
       [4.68000000e+02, 5.03000000e+02, 2.34000000e+02, 6.27000000e+02,
        1.42000000e+02],
       [1.83760684e+04, 7.75347913e+03, 1.15384615e+04, 3.34928230e+03,
        1.12676056e+04]])

None

(3, 5)

This numpy array has 3 rows and 5 columns!

Subsetting in higher dimensional numpy arrays

In higher dimensional numpy arrays, you can use subsetting as well. However, simply using square brackets with one entry will simply select rows. Let's see how it works.

first_row = None
first_row

array([8600000., 3900000., 2700000., 2100000., 1600000.])

Now how can we select a single element or a range of elements? There are two ways:

• you can either use two consecutive pairs of brackets, first one for row, second one for column
• you can also use a single pair of brackets and a comma as a seperator [row, column]

Now use these two methods consecutively to select the third element in the second row!

None

234.0

None

234.0

To select a full column, on the other hand, you can use a colon to denote that you want every element in the column. Select the population, size and density for the second city in the list (LA)

la_all = None
la_all

array([3.90000000e+06, 5.03000000e+02, 7.75347913e+03])

Sources

http://cs231n.github.io/python-numpy-tutorial/#numpy

http://www.numpy.org/

https://campus.datacamp.com/courses/intro-to-python-for-data-science/chapter-1-python-basics?ex=1

ds-numpy-lab's Issues

We will probably need to build this section out more

Python data science handbook is very good resource for a guide...intention is to explain the underlying functionality of the pandas library - for example universal functions are pretty key to pandas.