1
00:00:00,166 --> 00:00:02,566
Hello and welcome to this tutorial.

2
00:00:02,566 --> 00:00:05,300
Okay, so we are halfway
our data pre-processing phase.

3
00:00:05,300 --> 00:00:07,366
We learned how to import the libraries.

4
00:00:07,366 --> 00:00:08,733
To import the data set.

5
00:00:08,733 --> 00:00:10,866
We learned
how to take care of missing data.

6
00:00:10,866 --> 00:00:14,266
And today we're going to learn
how to encode categorical data.

7
00:00:14,833 --> 00:00:16,900
So the first thing that we're going to do

8
00:00:16,900 --> 00:00:20,233
is that we are going to explain
why we need to do this.

9
00:00:20,666 --> 00:00:23,666
So I'm going to go to Google Sheets here
to find my data set.

10
00:00:24,133 --> 00:00:28,300
And in this data sets we can see that
we have two categorical variables.

11
00:00:28,700 --> 00:00:31,933
We have the country variable here
and the purchase variable.

12
00:00:32,633 --> 00:00:35,133
These two variables
are categorical variables

13
00:00:35,133 --> 00:00:38,566
because simply
they contain categories. Here

14
00:00:38,566 --> 00:00:40,833
the country contains three categories.

15
00:00:40,833 --> 00:00:43,466
It's France, Spain and Germany.

16
00:00:43,466 --> 00:00:47,066
And the purchase variable contains
two categories yes and no.

17
00:00:47,500 --> 00:00:50,366
So that's
why they're called categorical variables.

18
00:00:50,366 --> 00:00:51,300
And now you can guess

19
00:00:51,300 --> 00:00:55,433
that since machine learning models
are based on mathematical equations,

20
00:00:55,700 --> 00:00:58,733
you can intuitively understand
that it would cause some problem

21
00:00:58,733 --> 00:01:01,733
if we keep the text here
and the categorical variables

22
00:01:01,766 --> 00:01:05,133
in the equations, because we would
only want numbers in the equations.

23
00:01:05,700 --> 00:01:09,966
So that's why we need to encode
the categorical variables,

24
00:01:09,966 --> 00:01:14,300
that is, to encode the text
that we have here into numbers.

25
00:01:15,533 --> 00:01:15,800
Okay.

26
00:01:15,800 --> 00:01:16,300
So in this

27
00:01:16,300 --> 00:01:20,066
tutorial we are going to encode these two
variables country and purchased.

28
00:01:20,333 --> 00:01:23,133
And you're going to see
that the technique is quite different.

29
00:01:23,133 --> 00:01:24,066
So here we are.

30
00:01:24,066 --> 00:01:27,066
We're going to use something
very practical in R

31
00:01:27,066 --> 00:01:28,566
that is the factor function.

32
00:01:28,566 --> 00:01:29,633
And the factor function

33
00:01:29,633 --> 00:01:33,933
will transform your categorical variables
into numeric categories.

34
00:01:34,233 --> 00:01:37,200
But it will see the variable as factors.

35
00:01:37,200 --> 00:01:40,200
And you will be able to choose the labels
of those factors.

36
00:01:40,366 --> 00:01:42,000
And due to the specific reason,

37
00:01:42,000 --> 00:01:45,000
we will need to create three columns
for each category.

38
00:01:45,166 --> 00:01:49,666
We will just transform the country column
into a column of factors,

39
00:01:50,100 --> 00:01:52,533
and we are going to specify
what the factors are.

40
00:01:52,533 --> 00:01:53,333
So let's do this.

41
00:01:53,333 --> 00:01:56,333
Let's start
by encoding the country column.

42
00:01:56,433 --> 00:01:58,333
So we just need to take
the column country.

43
00:01:58,333 --> 00:02:02,633
And to do this
we type data set dollar country

44
00:02:03,733 --> 00:02:04,333
equals.

45
00:02:04,333 --> 00:02:07,333
And then we use the factor
function factor.

46
00:02:08,133 --> 00:02:11,433
And in factor
we are going to specify three things.

47
00:02:11,433 --> 00:02:13,033
So let's have a look at the factor
function.

48
00:02:13,033 --> 00:02:15,700
To do this we click on F1 okay.

49
00:02:15,700 --> 00:02:18,866
So this contains
some info about the factor function.

50
00:02:18,866 --> 00:02:20,400
Let's look at the arguments.

51
00:02:20,400 --> 00:02:25,266
The first argument is the data
that we want to transform into factor.

52
00:02:25,300 --> 00:02:28,300
So this is of course going to be our.

53
00:02:30,033 --> 00:02:33,033
Country column from our data set okay.

54
00:02:33,566 --> 00:02:36,100
Then the second parameter is levels.

55
00:02:36,100 --> 00:02:39,633
And that's the names of the categories
in our country column.

56
00:02:40,266 --> 00:02:43,266
So let's add here levels equals.

57
00:02:43,500 --> 00:02:47,600
So are we going to write
the vector of levels here c parenthesis.

58
00:02:47,600 --> 00:02:49,900
And then we will input here
our three categories.

59
00:02:49,900 --> 00:02:50,933
So that's France.

60
00:02:52,933 --> 00:02:55,933
Spain and Germany.

61
00:02:56,700 --> 00:02:57,466
All right.

62
00:02:57,466 --> 00:03:00,466
So by the way C here is a vector and r.

63
00:03:00,733 --> 00:03:03,733
So by creating this c France Spain Germany

64
00:03:03,900 --> 00:03:06,900
we are creating a vector of three elements

65
00:03:07,400 --> 00:03:10,200
that is France Spain and Germany

66
00:03:10,200 --> 00:03:10,533
okay.

67
00:03:10,533 --> 00:03:11,866
And the last arguments

68
00:03:11,866 --> 00:03:15,900
that we need to input is labels
because we want to choose which number

69
00:03:16,000 --> 00:03:19,000
we want to give to France, to Spain
and to Germany.

70
00:03:19,066 --> 00:03:21,966
So we're going to add here labels.

71
00:03:21,966 --> 00:03:25,333
And then since we transform
our country categorical variable

72
00:03:25,333 --> 00:03:28,566
into factors,
we don't really care what numbers to use.

73
00:03:29,366 --> 00:03:32,133
So let's use one for France,

74
00:03:32,133 --> 00:03:35,466
two for Spain and three for Germany.

75
00:03:36,000 --> 00:03:38,500
So this has nothing to do
with my preference.

76
00:03:38,500 --> 00:03:40,366
It's not because I'm French
that I choose one.

77
00:03:40,366 --> 00:03:44,100
For France,
it is absolutely none order relative.

78
00:03:44,400 --> 00:03:47,400
So it's just by default that I use one,
two, three.

79
00:03:47,566 --> 00:03:50,233
Okay. So that's it.

80
00:03:50,233 --> 00:03:54,733
We just I just forgot one parenthesis
and I think this should disappear. Yes.

81
00:03:54,733 --> 00:03:55,466
Good.

82
00:03:55,466 --> 00:04:00,066
Okay, so the encoding of the country
categorical variable is ready.

83
00:04:00,400 --> 00:04:02,700
It's actually more simple than in Python.

84
00:04:02,700 --> 00:04:06,733
So before selecting and executing
this let's look at our data set.

85
00:04:06,866 --> 00:04:10,733
Our data
set contains our country written in text.

86
00:04:11,466 --> 00:04:12,900
And now let's select this

87
00:04:14,700 --> 00:04:16,000
execute.

88
00:04:16,000 --> 00:04:18,300
And now let's look at our data set.

89
00:04:18,300 --> 00:04:19,100
Perfect.

90
00:04:19,100 --> 00:04:21,866
Our country contains our encoded variable.

91
00:04:21,866 --> 00:04:26,766
And as you can see one is France,
two is Spain and three is Germany.

92
00:04:27,200 --> 00:04:28,333
Okay. So that's all good.

93
00:04:28,333 --> 00:04:31,333
And now we need to do the same
for the purchase column.

94
00:04:31,566 --> 00:04:32,700
And it's exactly the same.

95
00:04:32,700 --> 00:04:34,200
We're going to use the factor function.

96
00:04:34,200 --> 00:04:37,933
We are going to transform
the text categories into numerical labels.

97
00:04:38,300 --> 00:04:42,700
So let's do it right now
we're going to copy this paste it here.

98
00:04:43,066 --> 00:04:46,600
Here we're going to replace country
by purchased.

99
00:04:48,000 --> 00:04:49,633
All right.

100
00:04:49,633 --> 00:04:52,633
Same for here.

101
00:04:53,400 --> 00:04:56,166
Here we need to change the levels
because there are new categories.

102
00:04:56,166 --> 00:04:59,766
So the levels are going to be no.

103
00:05:01,866 --> 00:05:04,333
And yes okay.

104
00:05:04,333 --> 00:05:07,333
And now of course
we need to change the labels.

105
00:05:07,566 --> 00:05:10,733
So we're going to put zero for no

106
00:05:11,266 --> 00:05:14,266
and one for yes I think that's
what actually makes sense.

107
00:05:14,566 --> 00:05:17,633
And parentheses okay. And that's ready.

108
00:05:17,633 --> 00:05:20,400
Let's look at the data set.
The data set contains no.

109
00:05:20,400 --> 00:05:22,166
And yes in the purchase column.

110
00:05:22,166 --> 00:05:25,666
And now if I select this and execute

111
00:05:26,900 --> 00:05:29,600
the data set contains now zero and one.

112
00:05:29,600 --> 00:05:31,200
Perfect.

113
00:05:31,200 --> 00:05:32,100
We're all good.

114
00:05:32,100 --> 00:05:36,900
We, we encoded our categorical data in R,
and now you know everything

115
00:05:36,900 --> 00:05:41,700
that there is to know about encoding
categorical data in Python and R.

116
00:05:42,033 --> 00:05:43,466
So congratulations.

117
00:05:43,466 --> 00:05:47,166
We almost did the most difficult part
we are approaching soon.

118
00:05:47,166 --> 00:05:50,166
The fun and exciting part
about making models.

119
00:05:50,366 --> 00:05:51,666
So we are almost there.

120
00:05:51,666 --> 00:05:55,033
Hang on for 1 or 2 more tutorials
and this is going to get

121
00:05:55,033 --> 00:05:58,033
very exciting.