0
1
00:00:00,650 --> 00:00:05,960
In this lesson we're gonna be looking at variables again, but only this time these variables are going
1

2
00:00:05,960 --> 00:00:12,400
to be containing lots of data, instead of just a single number or a single piece of text.
2

3
00:00:12,440 --> 00:00:20,270
In other words, we're going to introduce you to collections. And by by collections we mean lists, arrays,
3

4
00:00:20,300 --> 00:00:22,640
data frames and series.
4

5
00:00:22,940 --> 00:00:29,270
In the last lesson we learned about how to store data inside variables and we learned about data types.
5

6
00:00:29,930 --> 00:00:31,910
In machine learning and data science,
6

7
00:00:31,940 --> 00:00:33,550
we often need to work with,
7

8
00:00:33,710 --> 00:00:35,460
well a lot of data.
8

9
00:00:35,630 --> 00:00:42,650
As such it would be silly to put each piece of data inside a separate variable.
9

10
00:00:42,800 --> 00:00:49,430
The alternative to assigning a single value to a single variable, like this, is that we have a single
10

11
00:00:49,430 --> 00:00:57,800
variable that actually contains lots of data and that variable will represent a collection. One of the
11

12
00:00:57,800 --> 00:01:02,960
most common types of collections that you'll encounter is the list.
12

13
00:01:02,960 --> 00:01:07,490
Lists are pre-built into Python and they're part of the language.
13

14
00:01:07,520 --> 00:01:09,950
Let me show you what a list looks like.
14

15
00:01:09,950 --> 00:01:19,280
I'm going to create a list called 'primeNumbers' and set it equal to a collection of 5 primes.
15

16
00:01:19,490 --> 00:01:23,670
And the way to do that is by using the square brackets syntax.
16

17
00:01:23,750 --> 00:01:32,650
I'm going to open a square bracket and then give my first value - 3, then put in a comma, give my second
17

18
00:01:32,650 --> 00:01:34,760
value - say 7,
18

19
00:01:34,990 --> 00:01:36,230
another comma,
19

20
00:01:36,460 --> 00:01:45,450
third value - 61, comma 29 and 199. And, just like that, we've created
20

21
00:01:45,540 --> 00:01:52,110
a variable called primeNumbers that holds onto a collection of five values.
21

22
00:01:52,140 --> 00:01:59,670
Now, just out of curiosity, let's take a look at what the type of the prime numbers variable is. 
22

23
00:01:59,670 --> 00:02:06,540
type(primeNumbers), Shift+Enter and we see that it's a list.
23

24
00:02:06,600 --> 00:02:12,020
Now one thing that I'll mention is that a list in Python, you know, doesn't have to be numbers, right.
24

25
00:02:12,140 --> 00:02:14,580
A list can contain all sorts of data.
25

26
00:02:14,730 --> 00:02:17,990
For example, here's what a list of strings would look like.
26

27
00:02:18,330 --> 00:02:22,830
coolPeople = []
27

28
00:02:22,830 --> 00:02:35,920
'Jay Z', 'Gandhi', 'me'
28

29
00:02:35,940 --> 00:02:36,720
There we go.
29

30
00:02:37,740 --> 00:02:43,560
The important thing to note is how we substituted Jay-Z for Kevin Spacey in this list.
30

31
00:02:43,630 --> 00:02:50,350
In any case, a list is just a variable that holds on to a bunch of values and it could even hold on to
31

32
00:02:50,350 --> 00:02:53,380
values of different types.
32

33
00:02:53,380 --> 00:03:00,470
Check this out: primesAndPeople = []
33

34
00:03:00,910 --> 00:03:12,340
'King Arthur', 17, 11, 'Jennifer Lopez'
34

35
00:03:12,340 --> 00:03:18,190
Now even though we've got three different lists; one with whole numbers, one with text and one with a
35

36
00:03:18,190 --> 00:03:24,550
mix of different types of data, all of these three variables are actually of the same type - they're of
36

37
00:03:24,550 --> 00:03:26,410
type list.
37

38
00:03:26,650 --> 00:03:31,750
So now that we've covered the syntax for creating a list, I'm going to show you something even cooler
38

39
00:03:31,810 --> 00:03:39,650
than storing data and lists - and that's retrieving data from a list. So the question is how do you pull
39

40
00:03:39,650 --> 00:03:43,420
out a particular value from one of these lists?
40

41
00:03:43,490 --> 00:03:44,920
Here's the secret.
41

42
00:03:45,110 --> 00:03:51,200
The way we grab a particular item from the list is by that item's index.
42

43
00:03:51,230 --> 00:03:53,220
What do I mean by index?
43

44
00:03:53,270 --> 00:03:56,660
The index is just the item's position in the list.
44

45
00:03:56,690 --> 00:03:58,220
Let me show you what I mean.
45

46
00:03:58,460 --> 00:04:07,730
If I type primeNumbers, and open the square brackets; between these two square brackets I can put a number
46

47
00:04:08,810 --> 00:04:11,950
and that number is the value of the index.
47

48
00:04:11,960 --> 00:04:20,150
So, say I put in the number two and hit Shift+Ente,r I get back the value 61, but 61 is the
48

49
00:04:20,150 --> 00:04:22,470
third item in this list.
49

50
00:04:22,670 --> 00:04:23,870
Why is it the third item?
50

51
00:04:24,440 --> 00:04:32,090
Well it turns out programmers really, really, really, really like to start counting from zero.
51

52
00:04:32,180 --> 00:04:40,310
In other words that first item, number 3, is at position 0, 7 is at position 1 and 61 is at position
52

53
00:04:40,310 --> 00:04:41,630
2.
53

54
00:04:41,720 --> 00:04:48,040
Now oftentimes what we're gonna be doing is storing the value that we pull out from a list in another
54

55
00:04:48,050 --> 00:04:50,960
variable - so we can do something like this"
55

56
00:04:51,200 --> 00:04:59,110
bestPrimeEver = primeNumbers[4]
56

57
00:04:59,260 --> 00:05:07,010
In this case we're pulling out one of the numbers from the list and storing it in a variable called
57

58
00:05:07,280 --> 00:05:09,180
bestPrimeEver.
58

59
00:05:09,340 --> 00:05:14,150
Now, I know you might have a different favorite prime number too, but I'm sorry to say it's not going
59

60
00:05:14,150 --> 00:05:16,520
to be as good as 199.
60

61
00:05:16,520 --> 00:05:18,940
Let me give you another example of a list.
61

62
00:05:18,980 --> 00:05:27,890
Say we had a list of eggs or egg objects and we want to pull out the egg at index 1, which egg
62

63
00:05:27,890 --> 00:05:29,390
would we be pulling out?
63

64
00:05:29,660 --> 00:05:33,760
Which egg is my egg equal to?
64

65
00:05:33,820 --> 00:05:39,230
The answer is we would be grabbing the second egg in the list.
65

66
00:05:39,550 --> 00:05:40,760
The spotted one.
66

67
00:05:41,080 --> 00:05:46,950
Again, this is because programmers start counting from zero.
67

68
00:05:47,080 --> 00:05:53,680
Now the reason I'm belaboring this point so much is because forgetting to count from zero is super common.
68

69
00:05:53,890 --> 00:05:58,850
And these off by 1 errors will also crash your Python program.
69

70
00:05:58,920 --> 00:06:07,130
Check out what happens when we write primeNumbers[5]  and hit Shift+Enter.
70

71
00:06:07,200 --> 00:06:15,430
We get an error that reads "index out of range" and that's because we're trying to select an item beyond
71

72
00:06:15,580 --> 00:06:17,230
the length of the list.
72

73
00:06:17,380 --> 00:06:28,480
Our list has five items and we're trying to select the item at position 0, 1, 2, 3, 4, 5. The item at position
73

74
00:06:28,480 --> 00:06:38,560
5 is the sixth one on the list. Since there is no sixth item, our Python program goes down in flames - "index
74

75
00:06:38,560 --> 00:06:40,000
out of range".
75

76
00:06:40,180 --> 00:06:48,140
So if you ever see these cursed words in your crash logs, now you know why. Let's change that five back
76

77
00:06:48,140 --> 00:06:55,950
to a four. Now lists are pretty common in Python and we'll be working with them a lot.
77

78
00:06:56,090 --> 00:07:02,510
However, we'll be encountering quite a few other kinds of collections too and a lot of them are very,
78

79
00:07:02,510 --> 00:07:10,760
very similar to the list but they are classified as a different data type because there are subtle differences.
79

80
00:07:11,780 --> 00:07:14,490
In scientific computing and machine learning,
80

81
00:07:14,510 --> 00:07:22,320
we're going to be working with another data structure very, very often - namely the array. An array is like
81

82
00:07:22,320 --> 00:07:23,850
the list's step brother.
82

83
00:07:23,950 --> 00:07:28,440
He's somewhat similar to its sister, but he's got a different dad.
83

84
00:07:28,440 --> 00:07:31,290
You see the Python list is like a ballerina.
84

85
00:07:31,290 --> 00:07:34,160
She's flexible, friendly, versatile.
85

86
00:07:34,350 --> 00:07:40,140
She can grow and shrink in size and she can hold onto different types of data like numbers and text
86

87
00:07:40,230 --> 00:07:45,010
in the same list. The array on the other hand is a weightlifter.
87

88
00:07:45,120 --> 00:07:50,430
He's strong and he's very good at particular tasks but he's less versatile.
88

89
00:07:50,430 --> 00:07:55,860
He's also so buff that if he tried to punch you he couldn't reach around his own pecs, but that's beside
89

90
00:07:55,860 --> 00:07:56,900
the point.
90

91
00:07:56,910 --> 00:08:03,090
All the data in an array must be of the same type, so an array must be all numbers or all strings,
91

92
00:08:03,090 --> 00:08:09,250
for example. Now, believe it or not, you've actually already encountered an array.
92

93
00:08:09,310 --> 00:08:15,730
Remember how we were looking at the regression intercept when we were estimating the movie revenue? Below
93

94
00:08:15,730 --> 00:08:16,390
the cell,
94

95
00:08:16,390 --> 00:08:24,340
we saw the output array([-7236000]).
95

96
00:08:25,140 --> 00:08:26,100
In this output,
96

97
00:08:26,110 --> 00:08:29,640
you even see the word array so we know that we're working with one.
97

98
00:08:29,890 --> 00:08:31,610
But let's make this more explicit.
98

99
00:08:31,630 --> 00:08:35,130
Let's check the type of 
99

100
00:08:35,290 --> 00:08:42,110
regr.intercept_ under school and hit Shift+Enter. Now, if you see this error,
100

101
00:08:42,130 --> 00:08:49,170
just like I do, it's because I haven't actually run any of the Python code in the prior cells so I'm
101

102
00:08:49,170 --> 00:08:57,040
gonna go to Cell > Run All and my notebook will know about the regression intercept so that now it will
102

103
00:08:57,100 --> 00:09:06,880
correctly evaluate this cell and what we discover is that the full name of this data type is 
103

104
00:09:07,240 --> 00:09:10,510
numpy.ndarray
104

105
00:09:10,510 --> 00:09:15,150
The cool thing is is that working with arrays is actually very similar to working with lists.
105

106
00:09:15,490 --> 00:09:18,520
In other words we can use that square brackets syntax.
106

107
00:09:18,520 --> 00:09:29,890
So if I write regr.intercept_[0] I'm pulling out the first and only
107

108
00:09:29,890 --> 00:09:33,140
value from our array. Hitting Shift+Enter,
108

109
00:09:33,190 --> 00:09:37,200
I see this value printed below the cell. Up here,
109

110
00:09:37,240 --> 00:09:42,910
in contrast, I was looking at the entire array, but in this cell I'm just looking at one of the values
110

111
00:09:43,120 --> 00:09:50,720
inside of this array and I'm accessing this value with the square brackets notation providing an index
111

112
00:09:51,020 --> 00:09:53,160
at value 0.
112

113
00:09:53,400 --> 00:10:00,980
Okay, so we learnt about lists and arrays in Python. Both of them use the square bracket notation and
113

114
00:10:00,980 --> 00:10:06,470
through both of them we can access individual items using an index.
114

115
00:10:06,680 --> 00:10:12,860
The index is just the position of the item in the list starting from zero.
115

116
00:10:12,860 --> 00:10:18,530
And this brings me to a very, very important point - namely do you know why the programmer quit his job?
116

117
00:10:20,420 --> 00:10:22,210
It's because he didn't get arrays...