0
1
00:00:00,550 --> 00:00:05,320
In this lesson we're going to wrap up our introduction to Python programming with something a little
1

2
00:00:05,320 --> 00:00:06,880
bit more fluffy.
2

3
00:00:07,300 --> 00:00:12,130
A hot topic on the Internet is always coding style.
3

4
00:00:12,490 --> 00:00:20,230
But what do I mean by coding style? Coding style is whether you use capital or lowercase letters, whether you
4

5
00:00:20,230 --> 00:00:24,520
use tabs or spaces, where you put your curly brackets and so on.
5

6
00:00:24,520 --> 00:00:31,930
It's how you arrange your code and what sort of syntax you use. Now the thing about programming style
6

7
00:00:32,020 --> 00:00:35,890
is that it's the classic bike shedding problem, right?
7

8
00:00:35,920 --> 00:00:39,010
Everybody has an opinion on what color the bike should be.
8

9
00:00:39,010 --> 00:00:44,530
And so everyone gets in a heated and lengthy argument about it in the same way.
9

10
00:00:44,530 --> 00:00:49,280
People also have very strong feelings on how code should be formatted.
10

11
00:00:49,390 --> 00:00:54,430
Now the irony is that the standards vary hugely between programming languages.
11

12
00:00:54,430 --> 00:01:00,880
So mainstream Python programming style is a little different from the mainstream Java programming style,
12

13
00:01:01,090 --> 00:01:04,690
which again is different from C++ or Swift.
13

14
00:01:05,650 --> 00:01:11,990
So the question is how do we find out what good Python programming style is or should be?
14

15
00:01:12,160 --> 00:01:18,160
And what are the guidelines that we should follow or what are the guidelines that people tend to follow?
15

16
00:01:18,160 --> 00:01:24,040
Well, lucky for us there are so-called style guides out there on the Internet.
16

17
00:01:24,040 --> 00:01:30,880
And the most popular style guide I think these days for Python is the official one from the Python Software
17

18
00:01:30,880 --> 00:01:32,070
Foundation.
18

19
00:01:32,260 --> 00:01:33,650
And it looks like this.
19

20
00:01:33,700 --> 00:01:37,790
It's called PEP 8 - the style guide for python code.
20

21
00:01:38,020 --> 00:01:45,780
And if you ask me this this makes excellent bedtime reading. Yeah, it's pretty long.
21

22
00:01:46,020 --> 00:01:54,810
It's very comprehensive and it's also not something that needs to be followed religiously.
22

23
00:01:54,810 --> 00:02:00,870
It's a suggestion for writing readable Python code, right? Readable for humans.
23

24
00:02:00,870 --> 00:02:04,440
The computers don't really care.
24

25
00:02:04,520 --> 00:02:08,730
The thing I'll say is that this isn't the only style guide out there.
25

26
00:02:08,730 --> 00:02:15,180
In fact many companies have their own style guides too, that, you know, might be slightly different from
26

27
00:02:15,180 --> 00:02:18,730
the ones of the Python Foundation for example.
27

28
00:02:18,750 --> 00:02:23,490
So here is the style guide from the engineers at Google.
28

29
00:02:23,490 --> 00:02:26,700
The Python style guide, to be specific.
29

30
00:02:26,700 --> 00:02:34,440
Now, I was actually looking through this and trying to spot a difference with the foundation's style
30

31
00:02:34,440 --> 00:02:42,870
guide and the thing I noticed is that these style guides are actually very, very similar and there are
31

32
00:02:42,970 --> 00:02:51,330
maybe like some differences; Python Software Foundation says line length should be 80 characters,
32

33
00:02:51,330 --> 00:02:55,860
Google says it should be 79.
33

34
00:02:56,040 --> 00:02:58,530
Most of the suggestions are actually really, really similar.
34

35
00:02:59,220 --> 00:03:04,940
So one of the differences I actually found seems to me more philosophical than anything else, right?
35

36
00:03:04,950 --> 00:03:08,620
So at the end Google says their parting words.
36

37
00:03:08,730 --> 00:03:10,070
BE CONSISTENT.
37

38
00:03:10,080 --> 00:03:10,260
Right?
38

39
00:03:10,260 --> 00:03:12,890
This is this is the formatting they use - all caps.
39

40
00:03:12,900 --> 00:03:14,490
Right?
40

41
00:03:14,850 --> 00:03:22,470
In contrast, the PEP 8 style guide from the Python Foundation says "A Foolish Consistency is the Hobgoblin
41

42
00:03:22,560 --> 00:03:25,040
of Little Minds".
42

43
00:03:25,500 --> 00:03:30,240
So what they're saying here is that you can be a bit more flexible in your coding style if you need
43

44
00:03:30,240 --> 00:03:31,200
to be.
44

45
00:03:31,230 --> 00:03:35,750
Now as I said before, in the end none of these style guides are hard and fast rules.
45

46
00:03:35,760 --> 00:03:39,710
They're more about aesthetics and readability of the code.
46

47
00:03:39,870 --> 00:03:41,730
So you yourself you can make a choice, right?
47

48
00:03:41,730 --> 00:03:45,600
You can choose one of these style guides and just stick to it.
48

49
00:03:45,630 --> 00:03:47,580
That's really all there is to it.
49

50
00:03:47,610 --> 00:03:53,370
That said, let's have a look that we've written so far in the lessons and see how it stacks up against
50

51
00:03:53,370 --> 00:03:57,330
the style guide from the Python Software Foundation.
51

52
00:03:57,330 --> 00:04:03,660
I'm not going to ask you to read through all of PEP 8 and play dpot the difference, but truth be told
52

53
00:04:03,810 --> 00:04:08,770
we've actually violated some of the style recommendations of PEP 8.
53

54
00:04:08,820 --> 00:04:15,060
So what I want to do in this video is kind of give an introduction to familiarize ourselves with what
54

55
00:04:15,060 --> 00:04:16,700
is considered good practice.
55

56
00:04:16,740 --> 00:04:23,560
And afterwards we can return to our Python code and make changes if necessary.
56

57
00:04:23,610 --> 00:04:29,070
One thing that we've already alluded to in the previous lessons is the indentation -
57

58
00:04:29,070 --> 00:04:37,840
So the layout of our code. And we recommended that we use four spaces for each indentation level.
58

59
00:04:38,070 --> 00:04:43,770
If we look back at our Jupyter notebook, we can see that this formatting has already been taken care
59

60
00:04:43,770 --> 00:04:45,590
of for us, right?
60

61
00:04:45,600 --> 00:04:52,860
So we can see that this print statement 'Open Door' is one, two, three, four spaces indented.
61

62
00:04:52,920 --> 00:04:59,890
So in many ways the editor that we're using - Jupyter notebook or other text editors - modern text editors,
62

63
00:05:00,060 --> 00:05:06,120
not Notepad actually shoulder some of this code formatting burden for us.
63

64
00:05:06,150 --> 00:05:10,090
Another thing that I want to point out is naming conventions.
64

65
00:05:10,200 --> 00:05:16,620
So I'm gonna dwell on this a little bit because it's very, very handy for reading and interpreting the
65

66
00:05:16,620 --> 00:05:19,070
code that other people have written.
66

67
00:05:19,170 --> 00:05:27,450
So naming conventions are handy for being able to read Python code and know a little bit more about
67

68
00:05:27,660 --> 00:05:34,900
what it is that you're looking at, just from how it's typed. So let's start with variables.
68

69
00:05:34,970 --> 00:05:41,630
Typically you'll see variables written as single letters, especially if you're doing math or physics
69

70
00:05:41,720 --> 00:05:43,440
or machine learning, right?
70

71
00:05:43,460 --> 00:05:51,740
So lowercase y, Capital X, these are variable names that you'll see very, very often. Now short variable
71

72
00:05:51,740 --> 00:05:57,230
names like this are good when there's like an accepted convention in place.
72

73
00:05:57,230 --> 00:06:03,170
So for example with scikit-learn, you'll often see the variable name for the training data as capital
73

74
00:06:03,230 --> 00:06:08,200
X and you'll see that target values as lower case y.
74

75
00:06:08,330 --> 00:06:13,490
And this is a convention that they adopt and try to be consistent with across, like, their whole code
75

76
00:06:13,490 --> 00:06:14,260
base.
76

77
00:06:14,510 --> 00:06:23,180
But there is also a big, big advantage from not using single letters for variable names because if you
77

78
00:06:23,180 --> 00:06:29,750
have a descriptive variable name, right, as opposed to a single letter you can help make your code a lot
78

79
00:06:29,840 --> 00:06:34,600
more readable. For a nice, long, descriptive variable name,
79

80
00:06:34,610 --> 00:06:40,370
you'll want to write it in all lowercase and you want to separate, if there's multiple words, you want
80

81
00:06:40,370 --> 00:06:46,760
to separate the words with an underscore, so lightning_mcqueen was an example of this kind
81

82
00:06:46,760 --> 00:06:53,720
of formatting. And if you really, really want to use a single letter variable names then please, please
82

83
00:06:53,720 --> 00:07:01,730
please do not use lowercase l or uppercase I or uppercase O as your variable name.
83

84
00:07:01,730 --> 00:07:07,970
It's really, really hard to tell these apart with smaller fonts, right,
84

85
00:07:07,970 --> 00:07:10,450
if you're looking at a screen. So yeah,
85

86
00:07:10,460 --> 00:07:18,140
single letter variable names, maybe just exclude these ones for readability. Now module and package
86

87
00:07:18,140 --> 00:07:22,670
names, they actually tend to follow the same naming conventions as variables really.
87

88
00:07:22,670 --> 00:07:22,870
Right?
88

89
00:07:22,880 --> 00:07:28,610
They're written in lowercase and the longer names have the words separated again with an underscore.
89

90
00:07:29,580 --> 00:07:35,860
And again, for functions and methods, we typically use lower case words separated by underscores as well.
90

91
00:07:36,910 --> 00:07:43,000
If the function in question has parameters, then the parameters tend to follow the same naming conventions
91

92
00:07:43,090 --> 00:07:45,390
as with variables.
92

93
00:07:45,460 --> 00:07:52,150
So with this in mind, we can actually quickly spot the coding style in our Python intro notebook where
93

94
00:07:52,150 --> 00:07:55,210
we've run afoul of these guidelines.
94

95
00:07:55,210 --> 00:08:01,570
For example, at the very top of our Python intro notebook we had some variables called myAge, restaurantBill
95

96
00:08:01,570 --> 00:08:06,970
and serviceCharge and these are written in camel case.
96

97
00:08:06,970 --> 00:08:13,060
So camel case means starting with a lowercase letter and then capitalizing each word.
97

98
00:08:13,150 --> 00:08:19,420
And this is quite a popular style in programming languages like Java but you see it much much less in
98

99
00:08:19,420 --> 00:08:21,830
Python. Cool.
99

100
00:08:21,840 --> 00:08:26,150
So that pretty much covers our introduction to Python programming.
100

101
00:08:26,160 --> 00:08:32,610
We'll be diving deeper into Python programming techniques and the subsequent lessons because programming
101

102
00:08:32,730 --> 00:08:37,390
is actually strongly linked to both data science and machine learning.
102

103
00:08:37,620 --> 00:08:44,190
And the reason is that programming goes well beyond how to call functions from scikit-learn. Knowing
103

104
00:08:44,190 --> 00:08:47,770
programming develops your mindset as an engineer.
104

105
00:08:48,030 --> 00:08:53,520
And to succeed in this course and with machine learning more generally, having an engineering mindset
105

106
00:08:53,700 --> 00:08:56,780
is key as part of this course.
106

107
00:08:56,820 --> 00:09:02,460
And in the workplace you're going to be putting a lot of different skills together from different fields.
107

108
00:09:02,640 --> 00:09:09,360
And those include mathematics, statistics, design and of course programming.
108

109
00:09:09,360 --> 00:09:14,400
And the reason is that outside of academia you're gonna be in a position where you're going to be trying
109

110
00:09:14,400 --> 00:09:22,560
to solve a real life business problem and figuring out how to solve this problem will test your engineering
110

111
00:09:22,560 --> 00:09:23,790
mindset.
111

112
00:09:23,980 --> 00:09:30,630
You know ,a common misconception these days is that to solve problems you need to use the most sophisticated
112

113
00:09:30,630 --> 00:09:34,910
model possible or, I don't know, artificial intelligence.
113

114
00:09:35,010 --> 00:09:37,050
Everybody is talking about A.I., right?
114

115
00:09:37,050 --> 00:09:38,910
Artificial intelligence.
115

116
00:09:38,910 --> 00:09:44,120
However, in reality A.I. is not the only thing that matters.
116

117
00:09:44,280 --> 00:09:47,700
And A.I. doesn't solve all problems.
117

118
00:09:47,700 --> 00:09:53,340
Now depending on your specific problem that you're faced with, A.I. might not even add all that much
118

119
00:09:53,340 --> 00:10:00,840
value, especially considering the high cost and the high amount of data required to build an A.I. system.
119

120
00:10:00,870 --> 00:10:06,960
So, you know, with your engineering mindset as an engineer you're going to be thinking about how to solve
120

121
00:10:06,960 --> 00:10:09,210
your problem in an efficient way.
121

122
00:10:09,210 --> 00:10:16,500
So maybe a simple model will do for your purposes and this is why this course will teach you a variety
122

123
00:10:16,500 --> 00:10:21,480
of tools and that way you can use the right tool for the job.
123

124
00:10:21,510 --> 00:10:27,310
After all, if all you've got is a hammer then everything starts to look like a nail.
124

125
00:10:27,420 --> 00:10:34,920
And coming back to programming, and programming in Python specifically, Python is actually a really, really
125

126
00:10:34,920 --> 00:10:44,160
nice language to start learning to program in contrast to other languages like C or C++, because Python
126

127
00:10:44,310 --> 00:10:50,560
actually takes care a lot of the error prone grunt work that goes on behind the scenes for you.
127

128
00:10:50,940 --> 00:10:56,970
And I don't think this is a bad thing because, unlike with C or C++, with Python you don't have to do
128

129
00:10:56,970 --> 00:11:03,810
chores like laying out the memory structure, managing the memory allocation or implementing access routines.
129

130
00:11:04,980 --> 00:11:05,980
Now, don't get me wrong,
130

131
00:11:06,050 --> 00:11:11,180
I think there is a time and place for all those things and becoming familiar with like how computer
131

132
00:11:11,180 --> 00:11:17,960
hardware actually works will definitely, definitely help down the line but I don't think it's useful
132

133
00:11:18,020 --> 00:11:23,510
to get confronted with this right away, especially if you've got so many other things to worry about.
133

134
00:11:24,290 --> 00:11:28,230
Which brings me to our next topic. Where are we going next?
134

135
00:11:28,340 --> 00:11:35,150
The next couple of lessons and sections will introduce you to the core of machine learning, namely loss
135

136
00:11:35,150 --> 00:11:39,320
functions and a technique called gradient descent.
136

137
00:11:39,350 --> 00:11:44,660
This is where the actual learning takes place and we're going to take a look at how this works under
137

138
00:11:44,660 --> 00:11:47,840
the hood for regression techniques.
138

139
00:11:47,960 --> 00:11:54,400
Also, we're going to be looking more closely at regressions that use more than one explanatory variable.
139

140
00:11:54,440 --> 00:11:57,910
In other words, we're gonna spice up our regression model.
140

141
00:11:57,980 --> 00:12:04,250
For example, if you were trying to predict a house price then you probably have to look at a number of
141

142
00:12:04,250 --> 00:12:11,480
different factors like the number of bedrooms, the location, the total area of the flat or the house, the
142

143
00:12:11,480 --> 00:12:14,970
proximity to amenities like a park or a good school.
143

144
00:12:14,970 --> 00:12:21,470
There's a number of factors that contribute to price of a house and the technique that will tease
144

145
00:12:21,470 --> 00:12:25,120
out the individual contribution of each of these factors.
145

146
00:12:25,220 --> 00:12:32,420
It's called multivariable regression and on that bombshell, it's time for me to get some more coffee
146

147
00:12:32,630 --> 00:12:35,150
and then I'll see you in the next lesson.
147

148
00:12:35,150 --> 00:12:36,080
Take care.