1
00:00:00,000 --> 00:00:04,067
[MUSIC]

2
00:00:04,067 --> 00:00:09,395
So next, we're going to inspect
some columns of our data set.

3
00:00:09,395 --> 00:00:15,921
So let me, let's inspect some

4
00:00:15,921 --> 00:00:21,195
columns of the data set.

5
00:00:28,145 --> 00:00:30,310
So let's take a look at it.

6
00:00:30,310 --> 00:00:34,950
So if I type sf['Country'],

7
00:00:34,950 --> 00:00:38,950
what it's gonna do is select
a column country, and

8
00:00:38,950 --> 00:00:42,060
since I don't do anything to it,
I don't set it equal to anything.

9
00:00:42,060 --> 00:00:43,350
It's just going to print it.

10
00:00:45,510 --> 00:00:49,580
Now, it says United States,
Canada, England, USA, Poland,

11
00:00:49,580 --> 00:00:51,930
United States, Switzerland.

12
00:00:51,930 --> 00:00:54,630
So, that's the whole column,
but printed horizontally.

13
00:00:55,830 --> 00:00:57,683
Now, let's look at another column.

14
00:00:57,683 --> 00:01:06,064
So, for example,
let's look at the column age, so, sf.age.

15
00:01:06,064 --> 00:01:08,095
I'm just gonna look at it.

16
00:01:08,095 --> 00:01:12,650
So I have, the column is 24,
23, 22, 23, 23, 22, 25.

17
00:01:12,650 --> 00:01:16,110
Same one we did a visualization up here.

18
00:01:16,110 --> 00:01:21,542
Now we can take the same column sf.age and
do some computations on it.

19
00:01:21,542 --> 00:01:26,776
So if you type .mean, it's not
saying that column is a mean column,

20
00:01:26,776 --> 00:01:31,210
but actually computing
the mean of that column.

21
00:01:31,210 --> 00:01:35,210
Which, the mean age is 23.142857,
and so on.

22
00:01:36,280 --> 00:01:39,976
Now, we can compute, for
example the max age.

23
00:01:39,976 --> 00:01:45,629
So, as of age, and let's see max.

24
00:01:45,629 --> 00:01:52,310
And again you get 25, that's maximum age.

25
00:01:53,550 --> 00:01:56,164
Now, this is some prequel stuff,
but let's do more.

26
00:01:56,164 --> 00:02:02,305
Let's create new columns.

27
00:02:06,554 --> 00:02:10,316
In our SFrame.

28
00:02:10,316 --> 00:02:16,670
So, when you're doing machine learning,
often taking columns

29
00:02:16,670 --> 00:02:20,910
and transforming them, creating new ones,
this is called feature engineering.

30
00:02:20,910 --> 00:02:22,412
We will do a bunch of those.

31
00:02:22,412 --> 00:02:24,914
I'm gonna do now,
a creation of a new column.

32
00:02:24,914 --> 00:02:27,118
So, let's go back and
print the SFrame, and

33
00:02:27,118 --> 00:02:30,930
you see that there's a First Name
column and there's a Last Name column.

34
00:02:30,930 --> 00:02:31,887
There's no full name column.

35
00:02:31,887 --> 00:02:34,951
Let's say you want to create a new
full name column, it's pretty easy.

36
00:02:34,951 --> 00:02:38,175
I wanna say sf,
our SFrame of Full Name, so

37
00:02:38,175 --> 00:02:41,590
there's gonna be a column
called Full Name.

38
00:02:42,930 --> 00:02:45,160
And what is that column equal to?

39
00:02:45,160 --> 00:02:48,049
Well, it's just a concatenation
of first name and last name.

40
00:02:48,049 --> 00:02:52,302
And in Python, it's really easy to take
strings and concatenate them together.

41
00:02:52,302 --> 00:02:53,591
All you have to do is put a plus sign.

42
00:02:53,591 --> 00:02:58,370
So we can do sf['Full Name'] is gonna be

43
00:02:58,370 --> 00:03:03,814
the First Name column, so ['First Name'],

44
00:03:03,814 --> 00:03:08,857
and then I'm gonna put a space between it,
so

45
00:03:08,857 --> 00:03:13,808
plus the space, plus the last name column.

46
00:03:13,808 --> 00:03:18,410
So, Last Name.

47
00:03:18,410 --> 00:03:19,850
Now, this is pretty cool.

48
00:03:19,850 --> 00:03:21,980
I just create a new feature, new column,

49
00:03:21,980 --> 00:03:25,310
called full name which is just
the first name space last name.

50
00:03:25,310 --> 00:03:27,840
It's very easy to do
these things in Python,

51
00:03:27,840 --> 00:03:29,590
especially if you don't have typos.

52
00:03:29,590 --> 00:03:32,920
I see that this is Fist Name
instead of First Name.

53
00:03:32,920 --> 00:03:33,981
So, and there you go.

54
00:03:33,981 --> 00:03:37,468
So, now if you go back and
show the SFrame again,

55
00:03:37,468 --> 00:03:43,380
you'll see there is a new column called
Full Name has been placed at the end.

56
00:03:43,380 --> 00:03:46,683
So Bob Smith is the first full name there.

57
00:03:46,683 --> 00:03:49,830
Now, you can do all sorts
of interesting things.

58
00:03:49,830 --> 00:03:54,392
So, for
example you could take the age column and

59
00:03:54,392 --> 00:03:57,850
I'm not gonna create a new problem.

60
00:03:57,850 --> 00:04:00,800
I'm just gonna do add 2
here to this age column.

61
00:04:00,800 --> 00:04:04,100
And you'll see that it's just
gonna add 2 to every element.

62
00:04:04,100 --> 00:04:06,030
And so,
you can do these kinds of operations.

63
00:04:06,030 --> 00:04:11,081
For example, you can do age and

64
00:04:11,081 --> 00:04:18,010
multiply it by the column age again.

65
00:04:18,010 --> 00:04:19,860
Now what happens when you do that?

66
00:04:19,860 --> 00:04:20,580
Age times age?

67
00:04:20,580 --> 00:04:23,262
You're gonna get age
squared in every element.

68
00:04:23,262 --> 00:04:25,653
And you get 556, 529 and so on.

69
00:04:25,653 --> 00:04:29,679
[MUSIC]