CoCalc -- introduction-to-numpy-video.ipynb

GitHub Repository: mrdbourke/zero-to-mastery-ml
Path: blob/master/section-2-data-science-and-ml-tools/introduction-to-numpy-video.ipynb
⁸⁷⁴ views

Kernel: Python 3

In [1]:

import numpy as np

1. DataTypes & Attributes

In [2]:

# NumPy's main datatype is ndarray
a1 = np.array([1, 2, 3])
a1

Out[2]:

array([1, 2, 3])

In [3]:

type(a1)

Out[3]:

numpy.ndarray

In [4]:

a2 = np.array([[1, 2.0, 3.3],
               [4, 5, 6.5]])

a3 = np.array([[[1, 2, 3],
                [4, 5, 6],
                [7, 8, 9]],
               [[10, 11, 12],
                [13, 14, 15],
                [16, 17, 18]]])

In [5]:

a2

Out[5]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [6]:

a3

Out[6]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [7]:

a1.shape

Out[7]:

(3,)

In [8]:

a2.shape

Out[8]:

(2, 3)

In [9]:

a3.shape

Out[9]:

(2, 3, 3)

In [12]:

a1

Out[12]:

array([1, 2, 3])

In [13]:

a2

Out[13]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [14]:

a3

Out[14]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [10]:

a1.ndim, a2.ndim, a3.ndim

Out[10]:

(1, 2, 3)

In [15]:

a1.dtype, a2.dtype, a3.dtype

Out[15]:

(dtype('int64'), dtype('float64'), dtype('int64'))

In [17]:

a3

Out[17]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [16]:

a1.size, a2.size, a3.size

Out[16]:

(3, 6, 18)

In [18]:

type(a1), type(a2), type(a3)

Out[18]:

(numpy.ndarray, numpy.ndarray, numpy.ndarray)

In [19]:

a2

Out[19]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [20]:

# Create a DataFrame from a NumPy array
import pandas as pd

df = pd.DataFrame(a2)
df

Out[20]:

2. Creating arrays

In [21]:

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

Out[21]:

array([1, 2, 3])

In [23]:

sample_array.dtype

Out[23]:

dtype('int64')

In [24]:

ones = np.ones((2, 3))

In [25]:

ones

Out[25]:

array([[1., 1., 1.],
       [1., 1., 1.]])

In [26]:

ones.dtype

Out[26]:

dtype('float64')

In [27]:

type(ones)

Out[27]:

numpy.ndarray

In [28]:

zeros = np.zeros((2, 3))

In [29]:

zeros

Out[29]:

array([[0., 0., 0.],
       [0., 0., 0.]])

In [30]:

range_array = np.arange(0, 10, 2)
range_array

Out[30]:

array([0, 2, 4, 6, 8])

In [32]:

random_array = np.random.randint(0, 10, size=(3, 5))
random_array

Out[32]:

array([[5, 9, 7, 6, 4],
       [1, 9, 6, 7, 1],
       [7, 5, 5, 9, 0]])

In [33]:

random_array.size

Out[33]:

15

In [34]:

random_array.shape

Out[34]:

(3, 5)

In [37]:

random_array_2 = np.random.random((5, 3))
random_array_2

Out[37]:

array([[0.22167345, 0.34437229, 0.73864985],
       [0.66670981, 0.82113975, 0.92073724],
       [0.03024207, 0.36297608, 0.39167642],
       [0.22192665, 0.85257304, 0.62190959],
       [0.15216592, 0.0638989 , 0.71405768]])

In [38]:

random_array_2.shape

Out[38]:

(5, 3)

In [122]:

random_array_3 = np.random.rand(5, 3)
random_array_3

Out[122]:

array([[0.72816128, 0.82122761, 0.76051512],
       [0.00714328, 0.42025683, 0.46313622],
       [0.0554995 , 0.54144213, 0.60777075],
       [0.82845319, 0.94180927, 0.12814785],
       [0.23043067, 0.6591584 , 0.13247399]])

In [98]:

# Pseudo-random numbers
np.random.seed(seed=99999)
random_array_4 = np.random.randint(10, size=(5, 3))
random_array_4

Out[98]:

array([[0, 3, 1],
       [8, 1, 3],
       [8, 5, 6],
       [0, 6, 0],
       [4, 0, 9]])

In [115]:

np.random.seed(7)
random_array_5 = np.random.random((5, 3))
random_array_5

Out[115]:

array([[0.07630829, 0.77991879, 0.43840923],
       [0.72346518, 0.97798951, 0.53849587],
       [0.50112046, 0.07205113, 0.26843898],
       [0.4998825 , 0.67923   , 0.80373904],
       [0.38094113, 0.06593635, 0.2881456 ]])

In [117]:

random_array_5 = np.random.random((5, 3))
random_array_5

Out[117]:

array([[0.20484909, 0.49076589, 0.37238469],
       [0.47740115, 0.36589039, 0.83791799],
       [0.76864751, 0.31399468, 0.57262533],
       [0.27604905, 0.45284293, 0.35297837],
       [0.65739946, 0.37035108, 0.45909298]])

In [123]:

random_array_4

Out[123]:

array([[0, 3, 1],
       [8, 1, 3],
       [8, 5, 6],
       [0, 6, 0],
       [4, 0, 9]])

3. Viewing arrays and matrices

In [124]:

np.unique(random_array_4)

Out[124]:

array([0, 1, 3, 4, 5, 6, 8, 9])

In [125]:

a1

Out[125]:

array([1, 2, 3])

In [126]:

a2

Out[126]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [127]:

a3

Out[127]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [128]:

a1[0]

Out[128]:

1

In [130]:

a2.shape

Out[130]:

(2, 3)

In [129]:

a2[0]

Out[129]:

array([1. , 2. , 3.3])

In [131]:

a3.shape

Out[131]:

(2, 3, 3)

In [132]:

a3[0]

Out[132]:

array([[1, 2, 3],
       [4, 5, 6],
       [7, 8, 9]])

In [133]:

a3

Out[133]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [134]:

a2

Out[134]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [135]:

a2[1]

Out[135]:

array([4. , 5. , 6.5])

In [136]:

a3

Out[136]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [140]:

a3.shape

Out[140]:

(2, 3, 3)

In [139]:

a3[:2, :2, :2]

Out[139]:

array([[[ 1,  2],
        [ 4,  5]],

       [[10, 11],
        [13, 14]]])

In [142]:

a4 = np.random.randint(10, size=(2, 3, 4, 5))
a4

Out[142]:

array([[[[8, 4, 9, 0, 2],
         [0, 7, 6, 2, 9],
         [9, 5, 1, 0, 0],
         [9, 1, 1, 5, 3]],

        [[2, 0, 4, 8, 7],
         [1, 4, 9, 3, 6],
         [7, 1, 0, 2, 7],
         [0, 0, 9, 8, 2]],

        [[0, 5, 5, 0, 9],
         [0, 8, 4, 5, 1],
         [4, 0, 7, 2, 0],
         [3, 9, 8, 3, 7]]],


       [[[1, 0, 1, 4, 6],
         [1, 0, 1, 0, 1],
         [8, 4, 1, 6, 8],
         [5, 2, 4, 7, 4]],

        [[1, 2, 8, 5, 2],
         [9, 6, 2, 5, 7],
         [7, 5, 5, 6, 0],
         [6, 4, 6, 2, 6]],

        [[8, 7, 0, 5, 8],
         [9, 7, 7, 6, 2],
         [5, 6, 1, 9, 3],
         [4, 0, 3, 5, 6]]]])

In [143]:

a4.shape, a4.ndim

Out[143]:

((2, 3, 4, 5), 4)

In [147]:

# Get the first 4 numbers of the inner most arrays
a4[:, :, :, :1]

Out[147]:

array([[[[8],
         [0],
         [9],
         [9]],

        [[2],
         [1],
         [7],
         [0]],

        [[0],
         [0],
         [4],
         [3]]],


       [[[1],
         [1],
         [8],
         [5]],

        [[1],
         [9],
         [7],
         [6]],

        [[8],
         [9],
         [5],
         [4]]]])

4. Manipulating & comparing arrays

Arithmetic

In [166]:

a1

Out[166]:

array([1, 2, 3])

In [168]:

ones = np.ones(3)
ones

Out[168]:

array([1., 1., 1.])

In [169]:

a1 + ones

Out[169]:

array([2., 3., 4.])

In [170]:

a1 - ones

Out[170]:

array([0., 1., 2.])

In [171]:

a1 * ones

Out[171]:

array([1., 2., 3.])

In [174]:

a1

Out[174]:

array([1, 2, 3])

In [172]:

a2

Out[172]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [173]:

a1 * a2

Out[173]:

array([[ 1. ,  4. ,  9.9],
       [ 4. , 10. , 19.5]])

In [176]:

a3

Out[176]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [177]:

# How can you reshape a2 to be compatible with a3? 
# Search: "How to reshape numpy array"
a2 * a3

Out[177]:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-177-37d8620f1b2c> in <module>
----> 1 a2 * a3

ValueError: operands could not be broadcast together with shapes (2,3) (2,3,3) 

In [178]:

a1 / ones

Out[178]:

array([1., 2., 3.])

In [180]:

a2 / a1

Out[180]:

array([[1.        , 1.        , 1.1       ],
       [4.        , 2.5       , 2.16666667]])

In [179]:

# Floor division removes the decimals (rounds down)
a2 // a1

Out[179]:

array([[1., 1., 1.],
       [4., 2., 2.]])

In [182]:

a2

Out[182]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [181]:

a2 ** 2

Out[181]:

array([[ 1.  ,  4.  , 10.89],
       [16.  , 25.  , 42.25]])

In [184]:

np.square(a2)

Out[184]:

array([[ 1.  ,  4.  , 10.89],
       [16.  , 25.  , 42.25]])

In [186]:

a1 + ones

Out[186]:

array([2., 3., 4.])

In [185]:

np.add(a1, ones)

Out[185]:

array([2., 3., 4.])

In [187]:

a1 % 2

Out[187]:

array([1, 0, 1])

In [188]:

a1 / 2

Out[188]:

array([0.5, 1. , 1.5])

In [189]:

a2 % 2

Out[189]:

array([[1. , 0. , 1.3],
       [0. , 1. , 0.5]])

In [190]:

np.exp(a1)

Out[190]:

array([ 2.71828183,  7.3890561 , 20.08553692])

In [191]:

np.log(a1)

Out[191]:

array([0.        , 0.69314718, 1.09861229])

Aggregation

Aggregation = performing the same operation on a number of things

In [193]:

listy_list = [1, 2, 3]
type(listy_list)

Out[193]:

list

In [194]:

sum(listy_list)

Out[194]:

6

In [197]:

sum(a1)

Out[197]:

6

In [198]:

np.sum(a1)

Out[198]:

6

Use Python's methods (sum()) on Python datatypes and use NumPy's methods on NumPy arrays (np.sum()).

In [199]:

# Creative a massive NumPy array
massive_array = np.random.random(100000)
massive_array.size

Out[199]:

100000

In [201]:

massive_array[:10]

Out[201]:

array([0.67959677, 0.44959741, 0.3896631 , 0.66184575, 0.01048353,
       0.26151906, 0.75223629, 0.93748825, 0.67568463, 0.49847177])

In [202]:

%timeit sum(massive_array) # Python's sum()
%timeit np.sum(massive_array) # NumPy's np.sum()

Out[202]:

17.9 ms ± 2.94 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)
34 µs ± 391 ns per loop (mean ± std. dev. of 7 runs, 10000 loops each)

In [203]:

17900 / 34

Out[203]:

526.4705882352941

In [204]:

a2

Out[204]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [205]:

np.mean(a2)

Out[205]:

3.6333333333333333

In [206]:

np.max(a2)

Out[206]:

6.5

In [207]:

np.min(a2)

Out[207]:

1.0

Standard deviation and variance are measures of 'spread' of data.

The higher standard deviation and the higher variance of data, the more spread out the values are.

The lower standard deviation and lower variance, the less spread out the values are.

In [208]:

# Standard deviation = a measure of how spread out a group of numbers is from the mean
np.std(a2)

Out[208]:

1.8226964152656422

In [209]:

# Variance = measure of the average degree to which each number is different to the mean
# Higher variance = wider range of numbers
# Lower variance = lower range of numbers
np.var(a2)

Out[209]:

3.3222222222222224

In [210]:

# Standard deviation = squareroot of variance
np.sqrt(np.var(a2))

Out[210]:

1.8226964152656422

In [211]:

# Demo of std and var
high_var_array = np.array([1, 100, 200, 300, 4000, 5000])
low_var_array = np.array([2, 4, 6, 8, 10])

In [212]:

np.var(high_var_array), np.var(low_var_array)

Out[212]:

(4296133.472222221, 8.0)

In [213]:

np.std(high_var_array), np.std(low_var_array)

Out[213]:

(2072.711623024829, 2.8284271247461903)

In [215]:

np.mean(high_var_array), np.mean(low_var_array)

Out[215]:

(1600.1666666666667, 6.0)

In [217]:

%matplotlib inline
import matplotlib.pyplot as plt
plt.hist(high_var_array)
plt.show()

Out[217]:

In [218]:

plt.hist(low_var_array)
plt.show()

Out[218]:

Reshaping & transposing

In [220]:

a2

Out[220]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [221]:

a2.shape

Out[221]:

(2, 3)

In [222]:

a3

Out[222]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [223]:

a3.shape

Out[223]:

(2, 3, 3)

In [224]:

a2 * a3

Out[224]:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-224-37d8620f1b2c> in <module>
----> 1 a2 * a3

ValueError: operands could not be broadcast together with shapes (2,3) (2,3,3) 

In [226]:

a2.shape

Out[226]:

(2, 3)

In [228]:

a2.reshape(2, 3, 1).shape

Out[228]:

(2, 3, 1)

In [229]:

a3.shape

Out[229]:

(2, 3, 3)

In [244]:

a2_reshape = a2.reshape(2, 3, 1)
a2_reshape

Out[244]:

array([[[1. ],
        [2. ],
        [3.3]],

       [[4. ],
        [5. ],
        [6.5]]])

In [231]:

a2_reshape * a3

Out[231]:

array([[[  1. ,   2. ,   3. ],
        [  8. ,  10. ,  12. ],
        [ 23.1,  26.4,  29.7]],

       [[ 40. ,  44. ,  48. ],
        [ 65. ,  70. ,  75. ],
        [104. , 110.5, 117. ]]])

In [233]:

a2

Out[233]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [235]:

a2.shape

Out[235]:

(2, 3)

In [232]:

# Transpose = switches the axis'
a2.T

Out[232]:

array([[1. , 4. ],
       [2. , 5. ],
       [3.3, 6.5]])

In [234]:

a2.T.shape

Out[234]:

(3, 2)

In [236]:

a3

Out[236]:

array([[[ 1,  2,  3],
        [ 4,  5,  6],
        [ 7,  8,  9]],

       [[10, 11, 12],
        [13, 14, 15],
        [16, 17, 18]]])

In [237]:

a3.shape

Out[237]:

(2, 3, 3)

In [239]:

a3.T

Out[239]:

array([[[ 1, 10],
        [ 4, 13],
        [ 7, 16]],

       [[ 2, 11],
        [ 5, 14],
        [ 8, 17]],

       [[ 3, 12],
        [ 6, 15],
        [ 9, 18]]])

In [240]:

a3.T.shape

Out[240]:

(3, 3, 2)

Dot product

In [256]:

np.random.seed(0)

mat1 = np.random.randint(10, size=(5, 3))
mat2 = np.random.randint(10, size=(5, 3))

mat1

Out[256]:

array([[5, 0, 3],
       [3, 7, 9],
       [3, 5, 2],
       [4, 7, 6],
       [8, 8, 1]])

In [257]:

mat2

Out[257]:

array([[6, 7, 7],
       [8, 1, 5],
       [9, 8, 9],
       [4, 3, 0],
       [3, 5, 0]])

In [258]:

mat1.shape, mat2.shape

Out[258]:

((5, 3), (5, 3))

In [260]:

mat1

Out[260]:

array([[5, 0, 3],
       [3, 7, 9],
       [3, 5, 2],
       [4, 7, 6],
       [8, 8, 1]])

In [261]:

mat2

Out[261]:

array([[6, 7, 7],
       [8, 1, 5],
       [9, 8, 9],
       [4, 3, 0],
       [3, 5, 0]])

In [259]:

# Element-wise multiplication (Hadamard product)
mat1 * mat2

Out[259]:

array([[30,  0, 21],
       [24,  7, 45],
       [27, 40, 18],
       [16, 21,  0],
       [24, 40,  0]])

In [262]:

# Dot product
np.dot(mat1, mat2)

Out[262]:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-262-0da3c94cc631> in <module>
      1 # Dot product
----> 2 np.dot(mat1, mat2)

<__array_function__ internals> in dot(*args, **kwargs)
ValueError: shapes (5,3) and (5,3) not aligned: 3 (dim 1) != 5 (dim 0)

In [263]:

# Transpose mat1
mat1.T

Out[263]:

array([[5, 3, 3, 4, 8],
       [0, 7, 5, 7, 8],
       [3, 9, 2, 6, 1]])

In [266]:

mat1.shape, mat2.T.shape

Out[266]:

((5, 3), (3, 5))

In [267]:

mat3 = np.dot(mat1, mat2.T)
mat3

Out[267]:

array([[ 51,  55,  72,  20,  15],
       [130,  76, 164,  33,  44],
       [ 67,  39,  85,  27,  34],
       [115,  69, 146,  37,  47],
       [111,  77, 145,  56,  64]])

In [268]:

mat3.shape

Out[268]:

(5, 5)

Dot product exmaple (nut butter sales)

In [269]:

np.random.seed(0)
# Number of jars sold
sales_amounts = np.random.randint(20, size=(5,3))
sales_amounts

Out[269]:

array([[12, 15,  0],
       [ 3,  3,  7],
       [ 9, 19, 18],
       [ 4,  6, 12],
       [ 1,  6,  7]])

In [270]:

# Create weekly_sales DataFrame
weekly_sales = pd.DataFrame(sales_amounts,
                            index=["Mon", "Tues", "Wed", "Thurs", "Fri"],
                            columns=["Almond butter", "Peanut butter", "Cashew butter"])
weekly_sales

Out[270]:

In [271]:

# Create prices array
prices = np.array([10, 8, 12])
prices

Out[271]:

array([10,  8, 12])

In [273]:

prices.shape

Out[273]:

(3,)

In [274]:

# Create butter_prices DataFrame
butter_prices = pd.DataFrame(prices.reshape(1, 3),
                             index=["Price"],
                             columns=["Almond butter", "Peanut butter", "Cashew butter"])
butter_prices

Out[274]:

In [279]:

prices.shape

Out[279]:

(3,)

In [278]:

sales_amounts.shape

Out[278]:

(5, 3)

In [276]:

total_sales = prices.dot(sales_amounts)

Out[276]:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-276-e4058314ac09> in <module>
----> 1 total_sales = prices.dot(sales_amounts)

ValueError: shapes (3,) and (5,3) not aligned: 3 (dim 0) != 5 (dim 0)

In [281]:

# Shape aren't aligned, let's transpose
total_sales = prices.dot(sales_amounts.T)
total_sales

Out[281]:

array([240, 138, 458, 232, 142])

In [284]:

# Create daily_sales
butter_prices.shape, weekly_sales.shape

Out[284]:

((1, 3), (5, 3))

In [286]:

weekly_sales.T.shape

Out[286]:

(3, 5)

In [292]:

daily_sales = butter_prices.dot(weekly_sales.T)
daily_sales.shape

Out[292]:

(1, 5)

In [293]:

weekly_sales.shape

Out[293]:

(5, 3)

In [288]:

weekly_sales

Out[288]:

In [291]:

# Doesn't work, not the right shape
weekly_sales["Total ($)"] = daily_sales
weekly_sales

Out[291]:

---------------------------------------------------------------------------
KeyError                                  Traceback (most recent call last)
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance)
   2896             try:
-> 2897                 return self._engine.get_loc(key)
   2898             except KeyError:
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
KeyError: 'Total ($)'

During handling of the above exception, another exception occurred:
KeyError                                  Traceback (most recent call last)
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/internals/managers.py in set(self, item, value)
   1068         try:
-> 1069             loc = self.items.get_loc(item)
   1070         except KeyError:
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance)
   2898             except KeyError:
-> 2899                 return self._engine.get_loc(self._maybe_cast_indexer(key))
   2900         indexer = self.get_indexer([key], method=method, tolerance=tolerance)
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
KeyError: 'Total ($)'

During handling of the above exception, another exception occurred:
ValueError                                Traceback (most recent call last)
<ipython-input-291-d761693dcb57> in <module>
----> 1 weekly_sales["Total ($)"] = daily_sales
      2 weekly_sales
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/frame.py in __setitem__(self, key, value)
   3470         else:
   3471             # set column
-> 3472             self._set_item(key, value)
   3473 
   3474     def _setitem_slice(self, key, value):
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/frame.py in _set_item(self, key, value)
   3548         self._ensure_valid_index(value)
   3549         value = self._sanitize_column(key, value)
-> 3550         NDFrame._set_item(self, key, value)
   3551 
   3552         # check if we are modifying a copy
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/generic.py in _set_item(self, key, value)
   3379 
   3380     def _set_item(self, key, value):
-> 3381         self._data.set(key, value)
   3382         self._clear_item_cache()
   3383 
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/internals/managers.py in set(self, item, value)
   1070         except KeyError:
   1071             # This item wasn't present, just insert at end
-> 1072             self.insert(len(self.items), item, value)
   1073             return
   1074 
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/internals/managers.py in insert(self, loc, item, value, allow_duplicates)
   1179         new_axis = self.items.insert(loc, item)
   1180 
-> 1181         block = make_block(values=value, ndim=self.ndim, placement=slice(loc, loc + 1))
   1182 
   1183         for blkno, count in _fast_count_smallints(self._blknos[loc:]):
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/internals/blocks.py in make_block(values, placement, klass, ndim, dtype, fastpath)
   3265         values = DatetimeArray._simple_new(values, dtype=dtype)
   3266 
-> 3267     return klass(values, ndim=ndim, placement=placement)
   3268 
   3269 
~/Desktop/ml-course/sample-project/env/lib/python3.7/site-packages/pandas/core/internals/blocks.py in __init__(self, values, placement, ndim)
    126             raise ValueError(
    127                 "Wrong number of items passed {val}, placement implies "
--> 128                 "{mgr}".format(val=len(self.values), mgr=len(self.mgr_locs))
    129             )
    130 
ValueError: Wrong number of items passed 5, placement implies 1

In [294]:

weekly_sales["Total ($)"] = daily_sales.T
weekly_sales

Out[294]:

Comparison Operators

In [295]:

a1

Out[295]:

array([1, 2, 3])

In [296]:

a2

Out[296]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [297]:

a1 > a2

Out[297]:

array([[False, False, False],
       [False, False, False]])

In [299]:

bool_array = a1 >= a2
bool_array

Out[299]:

array([[ True,  True, False],
       [False, False, False]])

In [301]:

type(bool_array), bool_array.dtype

Out[301]:

(numpy.ndarray, dtype('bool'))

In [302]:

a1 > 5

Out[302]:

array([False, False, False])

In [303]:

a1 < 5

Out[303]:

array([ True,  True,  True])

In [304]:

a1 == a1

Out[304]:

array([ True,  True,  True])

In [306]:

a1

Out[306]:

array([1, 2, 3])

In [307]:

a2

Out[307]:

array([[1. , 2. , 3.3],
       [4. , 5. , 6.5]])

In [305]:

a1 == a2

Out[305]:

array([[ True,  True, False],
       [False, False, False]])

5. Sorting arrays

In [310]:

random_array = np.random.randint(10, size=(3, 5))
random_array

Out[310]:

array([[7, 8, 1, 5, 9],
       [8, 9, 4, 3, 0],
       [3, 5, 0, 2, 3]])

In [311]:

random_array.shape

Out[311]:

(3, 5)

In [312]:

np.sort(random_array)

Out[312]:

array([[1, 5, 7, 8, 9],
       [0, 3, 4, 8, 9],
       [0, 2, 3, 3, 5]])

In [314]:

random_array

Out[314]:

array([[7, 8, 1, 5, 9],
       [8, 9, 4, 3, 0],
       [3, 5, 0, 2, 3]])

In [313]:

np.argsort(random_array)

Out[313]:

array([[2, 3, 0, 1, 4],
       [4, 3, 2, 0, 1],
       [2, 3, 0, 4, 1]])

In [315]:

a1

Out[315]:

array([1, 2, 3])

In [316]:

np.argsort(a1)

Out[316]:

array([0, 1, 2])

In [317]:

np.argmin(a1)

Out[317]:

0

In [318]:

np.argmax(a1)

Out[318]:

2

In [319]:

random_array

Out[319]:

array([[7, 8, 1, 5, 9],
       [8, 9, 4, 3, 0],
       [3, 5, 0, 2, 3]])

In [321]:

np.argmax(random_array, axis=0)

Out[321]:

array([1, 1, 1, 0, 0])

In [322]:

np.argmax(random_array, axis=1)

Out[322]:

array([4, 1, 1])

6. Practical Example - NumPy in Action!!!!

In [325]:

# Turn an image into a NumPy array
from matplotlib.image import imread

panda = imread("images/panda.png")
print(type(panda))

Out[325]:

<class 'numpy.ndarray'>

In [327]:

panda.size, panda.shape, panda.ndim

Out[327]:

(24465000, (2330, 3500, 3), 3)

In [330]:

panda[:5]

Out[330]:

array([[[0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        ...,
        [0.16470589, 0.12941177, 0.09411765],
        [0.16470589, 0.12941177, 0.09411765],
        [0.16470589, 0.12941177, 0.09411765]],

       [[0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        ...,
        [0.16470589, 0.12941177, 0.09411765],
        [0.16470589, 0.12941177, 0.09411765],
        [0.16470589, 0.12941177, 0.09411765]],

       [[0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        ...,
        [0.16470589, 0.12941177, 0.09411765],
        [0.16470589, 0.12941177, 0.09411765],
        [0.16470589, 0.12941177, 0.09411765]],

       [[0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        ...,
        [0.16862746, 0.13333334, 0.09803922],
        [0.16862746, 0.13333334, 0.09803922],
        [0.16862746, 0.13333334, 0.09803922]],

       [[0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        [0.05490196, 0.10588235, 0.06666667],
        ...,
        [0.16862746, 0.13333334, 0.09803922],
        [0.16862746, 0.13333334, 0.09803922],
        [0.16862746, 0.13333334, 0.09803922]]], dtype=float32)

In [331]:

car = imread("images/car-photo.png")
print(type(car))

Out[331]:

<class 'numpy.ndarray'>

In [333]:

car[:1]

Out[333]:

array([[[0.5019608 , 0.50980395, 0.4862745 , 1.        ],
        [0.3372549 , 0.34509805, 0.30588236, 1.        ],
        [0.20392157, 0.21568628, 0.14901961, 1.        ],
        ...,
        [0.64705884, 0.7058824 , 0.54901963, 1.        ],
        [0.59607846, 0.63529414, 0.45882353, 1.        ],
        [0.44705883, 0.47058824, 0.3372549 , 1.        ]]], dtype=float32)

In [334]:

dog = imread("images/dog-photo.png")
print(type(dog))

Out[334]:

<class 'numpy.ndarray'>

In [335]:

dog

Out[335]:

array([[[0.70980394, 0.80784315, 0.88235295, 1.        ],
        [0.72156864, 0.8117647 , 0.8862745 , 1.        ],
        [0.7411765 , 0.8156863 , 0.8862745 , 1.        ],
        ...,
        [0.49803922, 0.6862745 , 0.8392157 , 1.        ],
        [0.49411765, 0.68235296, 0.8392157 , 1.        ],
        [0.49411765, 0.68235296, 0.8352941 , 1.        ]],

       [[0.69411767, 0.8039216 , 0.8862745 , 1.        ],
        [0.7019608 , 0.8039216 , 0.88235295, 1.        ],
        [0.7058824 , 0.80784315, 0.88235295, 1.        ],
        ...,
        [0.5019608 , 0.6862745 , 0.84705883, 1.        ],
        [0.49411765, 0.68235296, 0.84313726, 1.        ],
        [0.49411765, 0.68235296, 0.8392157 , 1.        ]],

       [[0.6901961 , 0.8       , 0.88235295, 1.        ],
        [0.69803923, 0.8039216 , 0.88235295, 1.        ],
        [0.7058824 , 0.80784315, 0.88235295, 1.        ],
        ...,
        [0.5019608 , 0.6862745 , 0.84705883, 1.        ],
        [0.49803922, 0.6862745 , 0.84313726, 1.        ],
        [0.49803922, 0.6862745 , 0.84313726, 1.        ]],

       ...,

       [[0.9098039 , 0.81960785, 0.654902  , 1.        ],
        [0.8352941 , 0.7490196 , 0.6509804 , 1.        ],
        [0.72156864, 0.6313726 , 0.5372549 , 1.        ],
        ...,
        [0.01568628, 0.07058824, 0.02352941, 1.        ],
        [0.03921569, 0.09411765, 0.03529412, 1.        ],
        [0.03921569, 0.09019608, 0.05490196, 1.        ]],

       [[0.9137255 , 0.83137256, 0.6784314 , 1.        ],
        [0.8117647 , 0.7294118 , 0.627451  , 1.        ],
        [0.65882355, 0.5686275 , 0.47843137, 1.        ],
        ...,
        [0.00392157, 0.05490196, 0.03529412, 1.        ],
        [0.03137255, 0.09019608, 0.05490196, 1.        ],
        [0.04705882, 0.10588235, 0.06666667, 1.        ]],

       [[0.9137255 , 0.83137256, 0.68235296, 1.        ],
        [0.76862746, 0.68235296, 0.5882353 , 1.        ],
        [0.59607846, 0.5058824 , 0.44313726, 1.        ],
        ...,
        [0.03921569, 0.10196079, 0.07058824, 1.        ],
        [0.02745098, 0.08235294, 0.05882353, 1.        ],
        [0.05098039, 0.11372549, 0.07058824, 1.        ]]], dtype=float32)

In [ ]: