1
00:00:00,410 --> 00:00:04,680
All right now we've had a little bit of a conceptual overview of num pie.

2
00:00:04,710 --> 00:00:08,130
Let's get hands on and start to write some num pi code.

3
00:00:08,890 --> 00:00:13,620
So I'm going to follow the little workflow that we've we've kind of set up over the past few lessons

4
00:00:13,620 --> 00:00:16,790
and sections start off in a terminal.

5
00:00:16,910 --> 00:00:21,620
And I'm going to change directory into this little folder that I've got on my desktop.

6
00:00:21,620 --> 00:00:26,290
Yours may be stored somewhere else to see the slash airmail course.

7
00:00:26,390 --> 00:00:30,650
Now I've got a folder in here called sample project which has my environment.

8
00:00:30,650 --> 00:00:34,480
You'll want to change into the same folder that contains your environment.

9
00:00:35,390 --> 00:00:36,210
Hit enter.

10
00:00:36,220 --> 00:00:38,130
Now nowhere in the sample project folder.

11
00:00:38,150 --> 00:00:39,110
Beautiful.

12
00:00:39,660 --> 00:00:46,580
And we're going to check using Condit end list where our environments are so I can say I've got one

13
00:00:46,580 --> 00:00:48,170
here I've got a few currently.

14
00:00:48,420 --> 00:00:54,170
Base is activated because got the star but I want this one because that's in my sample project folder

15
00:00:54,510 --> 00:00:55,630
so I'm going to type it out.

16
00:00:55,640 --> 00:01:01,340
You could copy this and do Conda activate this here but where in the practice of writing things for

17
00:01:01,340 --> 00:01:09,140
ourselves because I want to get as used to as possible writing as much code as possible email cause

18
00:01:10,340 --> 00:01:11,560
sample project.

19
00:01:11,560 --> 00:01:17,020
Now remember you'll want to change yours to wherever your environment folder is stored activate the

20
00:01:17,020 --> 00:01:23,680
first thing we notice is that this has changed from base to our active environment now and next step

21
00:01:23,770 --> 00:01:31,010
is to create a Jupiter notebook our favorite workspace tool let's load that up now listen open our browser

22
00:01:32,090 --> 00:01:37,040
we brought to the Jupiter dashboard it's nice and zoomed in excellent we've got our folders here and

23
00:01:37,040 --> 00:01:40,030
files that are within our sample project folder.

24
00:01:40,070 --> 00:01:45,140
Yours may be a little bit different but we can see in my case about the introduction to Panda's notebook

25
00:01:45,170 --> 00:01:50,300
the example notebook from before my environment folder we're not worried about these we're going to

26
00:01:50,600 --> 00:01:58,250
start a new notebook Python 3 first thing we're gonna do is rename it to introduction to num PI because

27
00:01:58,250 --> 00:01:59,630
that's what we're working on.

28
00:01:59,780 --> 00:02:06,020
Hit enter and we'll save that when we come back to the dashboard we've now got introduction to num pi

29
00:02:06,140 --> 00:02:08,020
running beautiful.

30
00:02:08,390 --> 00:02:17,660
Now the first step for using num pi is to import it import num pi as an P MP is just short for num pi

31
00:02:18,050 --> 00:02:24,170
if you see this abbreviation somewhere you can probably safely assume that it's referring to num pi

32
00:02:24,290 --> 00:02:30,080
it's just to prevent us from having to type out num pi every single time we want to use a command we'll

33
00:02:30,110 --> 00:02:36,050
hit shift and enter beautiful now in this section and we're going to be focused on num PIs data type

34
00:02:36,350 --> 00:02:42,050
and attributes so that's all we'll create a little heading because we are in the business of making

35
00:02:42,050 --> 00:02:49,160
attractive and communicate of notebooks data types and attributes will escape and M to turn it into

36
00:02:49,160 --> 00:02:53,380
a markdown cell wonderful data types and attributes.

37
00:02:53,390 --> 00:03:06,030
Now data types as a plural is kind of a lie because none pi main data type is indeed array it's only

38
00:03:06,030 --> 00:03:12,720
really got one day at a time so everything you see in num pi is likely an NDA array and this stands

39
00:03:12,720 --> 00:03:20,940
for n dimensional array and an array can be almost any list of numbers you can imagine an n dimensions

40
00:03:20,940 --> 00:03:25,770
means or the list of numbers could be almost any shape that you can imagine but rather than talk about

41
00:03:25,770 --> 00:03:33,690
it let's create it so A1 for array one will create a sample array now you can do that by going num pi

42
00:03:33,840 --> 00:03:42,120
dot array which just means access the array function from num Pi which we've imported as MP and to build

43
00:03:42,120 --> 00:03:48,510
an array which is a list of numbers we want to pass at a list of numbers so let's do that and then to

44
00:03:48,510 --> 00:03:53,140
see it we'll type it again shift and enter Excellent.

45
00:03:53,190 --> 00:03:56,700
We've got our first name pie right now this is exciting.

46
00:03:56,700 --> 00:04:01,870
Let's check the top of it nun pie DOT and the array.

47
00:04:01,950 --> 00:04:02,520
Wonderful.

48
00:04:02,520 --> 00:04:04,110
So that's the same as here.

49
00:04:04,170 --> 00:04:07,000
Remember this is just n dimensional array.

50
00:04:07,140 --> 00:04:14,010
Now for some more examples let's create a few more arrays empty array or type out this one.

51
00:04:14,100 --> 00:04:16,240
This will have two brackets.

52
00:04:16,290 --> 00:04:18,390
We'll see why in a second.

53
00:04:18,390 --> 00:04:19,580
Now we'll separate it.

54
00:04:19,740 --> 00:04:20,760
We'll go to another line.

55
00:04:20,760 --> 00:04:23,890
One two three four five actually.

56
00:04:24,030 --> 00:04:26,070
Let's make this six point five.

57
00:04:26,070 --> 00:04:28,150
And this three point three.

58
00:04:28,200 --> 00:04:32,690
So we've got some floats as well as integers inside num pi.

59
00:04:32,700 --> 00:04:39,540
There we go we might make another one called a three empty array and this one we might actually do three

60
00:04:39,540 --> 00:04:40,440
brackets.

61
00:04:40,440 --> 00:04:42,950
So we're getting a bit adventurous here.

62
00:04:43,050 --> 00:04:45,370
We'll make this one two three.

63
00:04:45,390 --> 00:04:47,410
Let's go all the way up to 18.

64
00:04:47,490 --> 00:04:50,640
Bear with me type along if you can.

65
00:04:50,700 --> 00:04:56,110
Five six and then we'll go seven eight.

66
00:04:56,130 --> 00:04:57,880
Want to make sure it looks all right.

67
00:04:57,910 --> 00:05:00,790
Nine and then leave two brackets here.

68
00:05:00,810 --> 00:05:01,740
Oh actually yeah.

69
00:05:01,740 --> 00:05:02,520
Two brackets.

70
00:05:02,520 --> 00:05:03,960
That's correct.

71
00:05:03,960 --> 00:05:10,980
I'm just counting the number of brackets you'll see why this is important in a moment which is typing

72
00:05:10,980 --> 00:05:18,990
out the numbers from one to 18 in a very kind of cryptic format at the moment 13 14 15.

73
00:05:19,060 --> 00:05:22,980
Got a little bit more 16 17 18.

74
00:05:22,980 --> 00:05:26,370
My goodness we made it in the end.

75
00:05:26,370 --> 00:05:27,640
Let's see if this cell runs.

76
00:05:27,660 --> 00:05:28,330
Beautiful.

77
00:05:28,350 --> 00:05:32,250
If in doubt run the code now why would we do that.

78
00:05:32,250 --> 00:05:33,920
Well let's have a look at these two arrays.

79
00:05:33,990 --> 00:05:39,200
A two there we go we can see that we've got two sets of brackets here.

80
00:05:39,290 --> 00:05:40,820
You got two more here.

81
00:05:40,840 --> 00:05:44,200
Now let's have a look at a three.

82
00:05:44,620 --> 00:05:45,190
OK.

83
00:05:45,250 --> 00:05:49,810
Now this one it's kind of come out the exact same way as we've got here.

84
00:05:49,810 --> 00:05:50,630
Mm hmm.

85
00:05:50,650 --> 00:05:55,960
Now before we get any further and check out some of the attributes that these arrays have built in let's

86
00:05:55,960 --> 00:05:58,930
have a look at the anatomy of an array.

87
00:05:58,960 --> 00:06:01,590
So this is the anatomy of a name pi array.

88
00:06:01,600 --> 00:06:04,040
This is a one the one that we've just created.

89
00:06:04,060 --> 00:06:05,240
This is a two.

90
00:06:05,350 --> 00:06:09,410
And this is a three as a reason why I put in these numbers here.

91
00:06:09,430 --> 00:06:13,470
So we can exemplify the kinds of shapes of different arrays.

92
00:06:13,470 --> 00:06:18,340
Now this is very important in machine learning because oftentimes with a machine learning algorithm

93
00:06:18,370 --> 00:06:25,300
your main focus will be making sure the shape of your data that goes into a machine learning algorithm

94
00:06:25,600 --> 00:06:29,940
lines up with the other data that you're putting in as well as the output.

95
00:06:30,140 --> 00:06:32,230
But we'll see that in future sections.

96
00:06:32,230 --> 00:06:35,780
This is just learning the shape of different num pi arrays.

97
00:06:35,830 --> 00:06:37,350
So I've got one here.

98
00:06:37,420 --> 00:06:39,720
There's one row and three columns.

99
00:06:40,150 --> 00:06:43,030
So this would be shape 1 3.

100
00:06:43,030 --> 00:06:50,400
Now this kind of data here is referred to as an array or vector a vector is just a one dimensional array.

101
00:06:50,790 --> 00:06:56,230
If we go to this next one which is A2 we've got two rows and three columns.

102
00:06:56,230 --> 00:07:03,070
So this one has the shape to three it's more than one dimension and can be referred to as a matrix.

103
00:07:03,100 --> 00:07:07,370
So if you hear something referred to as a matrix it's probably got more than one dimension.

104
00:07:07,380 --> 00:07:09,330
So see how a vector vectors just one row.

105
00:07:09,370 --> 00:07:11,530
This matrix has two rows.

106
00:07:11,530 --> 00:07:17,020
And if we go down to a three we're starting to get three dimensional here because we've got got the

107
00:07:17,020 --> 00:07:18,640
access going back here.

108
00:07:18,700 --> 00:07:22,570
This is where the N N N dimensional array comes from.

109
00:07:22,570 --> 00:07:28,210
That's very hard for us humans to visualize anything more than three dimensions.

110
00:07:28,210 --> 00:07:33,280
But in num pi and with computers you could have as many ends as you want.

111
00:07:33,280 --> 00:07:38,590
So this could go up to four to five up to six up to seven depending on what type of data are working

112
00:07:38,590 --> 00:07:39,040
with.

113
00:07:39,040 --> 00:07:45,720
You may have thousands of different dimensions but if you hear axis zero it's usually referring to rows.

114
00:07:45,790 --> 00:07:52,960
Axis one is usually referring to columns and then access n is usually referring to dimensions that are

115
00:07:52,960 --> 00:07:54,970
beyond just rows and columns.

116
00:07:55,600 --> 00:08:04,030
So this array here or A3 would have the shape to three three because there's two rows three columns

117
00:08:04,410 --> 00:08:06,430
and we've got three across here.

118
00:08:06,430 --> 00:08:07,790
So that's the final one here.

119
00:08:07,810 --> 00:08:13,280
And now if n was higher there might be more commas and this little shape would extend out here.

120
00:08:13,840 --> 00:08:14,500
Okay.

121
00:08:14,740 --> 00:08:18,380
So now we've got the anatomy of a no higher rate down pat.

122
00:08:18,430 --> 00:08:21,360
Don't worry you can take a screenshot of this if you like.

123
00:08:21,370 --> 00:08:24,100
Otherwise it'll be in the resources section.

124
00:08:24,160 --> 00:08:29,170
Let's have a look at some of the attributes associated with no higher res.

125
00:08:29,170 --> 00:08:38,460
So the first one we'll take a look at is a one dot shape so it's three comma something.

126
00:08:38,460 --> 00:08:39,900
Now this is blank.

127
00:08:40,050 --> 00:08:48,900
Now the way num pi kind of format this data is that although we've said that the shape is 1 3 num pi

128
00:08:48,930 --> 00:08:52,080
kind of stores it without the row at all.

129
00:08:52,080 --> 00:08:54,900
So it's just got the three numbers here.

130
00:08:54,900 --> 00:08:58,140
That's just a little tidbit of how num pi displays some kind of shapes.

131
00:08:58,140 --> 00:09:04,140
But if we have a look at the shape of A2 we can see that it's the same as what we've used in our anatomy

132
00:09:04,140 --> 00:09:05,610
of a known pi array.

133
00:09:05,680 --> 00:09:12,740
Then if we go the same for a 3 2 3 3 OK let's have a look at a few more.

134
00:09:12,800 --> 00:09:20,300
So if we are a one and dim which is short for number of dimensions maybe we do a two dot end them and

135
00:09:20,360 --> 00:09:21,860
a three them.

136
00:09:22,370 --> 00:09:29,900
And what putting the comma here will do is when the code comes out it'll just print it with commas spacing

137
00:09:29,900 --> 00:09:31,890
that three different variables apart.

138
00:09:31,910 --> 00:09:33,800
So this is number of dimensions.

139
00:09:33,800 --> 00:09:36,690
So A1 let's have a look at it again.

140
00:09:36,860 --> 00:09:39,190
One has a dimension of 1.

141
00:09:39,320 --> 00:09:49,660
We'll have a look at A2 as a dimension of 2 and A3 has three dimensions.

142
00:09:49,710 --> 00:09:55,870
And so this is where the MDMA data type comes in is that these dimensions depending on how complex your

143
00:09:55,870 --> 00:09:58,330
data is could just keep going.

144
00:09:58,360 --> 00:10:02,710
And now the whole premise here is we'll see some practical examples in the next couple of videos is

145
00:10:02,710 --> 00:10:09,010
that whatever data you're working with it somehow gets turned into numbers and then a machine learning

146
00:10:09,010 --> 00:10:12,250
algorithm will find patterns in these numbers.

147
00:10:12,250 --> 00:10:17,110
Now depending on what data you're working with will depend on how many dimensions Zehra.

148
00:10:17,270 --> 00:10:18,390
But let's not stop there.

149
00:10:18,400 --> 00:10:21,270
Let's have a look at the D type attribute.

150
00:10:21,310 --> 00:10:23,990
You might be able to guess what data type these are.

151
00:10:24,040 --> 00:10:33,300
Let's go here shift into so we can see in our A1 array we've got ints in our A2 we've got floats as

152
00:10:33,300 --> 00:10:34,120
well as int.

153
00:10:34,170 --> 00:10:39,690
But it's going to default to float because that can only be one kind of data type so it will automatically

154
00:10:39,690 --> 00:10:48,230
just turn the rest of the numbers in here into floats and A3 is also just integers we've got a couple

155
00:10:48,230 --> 00:10:48,580
more.

156
00:10:48,650 --> 00:10:59,570
A one size a two size a three size we've got three elements in a one we've got six elements in a two

157
00:10:59,600 --> 00:11:02,870
and we've got 18 elements in a three.

158
00:11:03,080 --> 00:11:10,730
If we have a look at a three that makes sense beautiful so the size attribute tells us how many elements

159
00:11:10,730 --> 00:11:13,800
we've got total in our right now.

160
00:11:13,970 --> 00:11:20,680
One last one which is type 2 type 2 type a 3.

161
00:11:20,900 --> 00:11:24,470
I think we've already seen this but I'm going to get you to guess what it is before we hit shift and

162
00:11:24,550 --> 00:11:29,850
into 3 2 1 these are all in the array.

163
00:11:29,860 --> 00:11:38,140
So although they vary in size and shape a A1 a two and a three are all of the type NDA right now this

164
00:11:38,140 --> 00:11:40,640
is an important concept to remember.

165
00:11:41,080 --> 00:11:47,980
No matter what type or what shape your number higher rate is all of the functions and different data

166
00:11:47,980 --> 00:11:50,920
manipulations that you see throughout this section.

167
00:11:50,920 --> 00:11:52,380
If it's in an umpire.

168
00:11:52,390 --> 00:11:58,900
If your data is in a m the array data type all of the things were about to go through can be used on

169
00:11:58,900 --> 00:11:59,670
that.

170
00:11:59,710 --> 00:12:06,520
So that's a very important point to remember is that the NDA Ray is a universal data type and it's very

171
00:12:06,580 --> 00:12:14,380
very flexible to now to really exemplify num pi being the backbone of other scientific data packages

172
00:12:14,380 --> 00:12:16,050
that is let's put a note here.

173
00:12:16,060 --> 00:12:22,250
Create a data frame from num pi array.

174
00:12:22,890 --> 00:12:29,410
So let's have a look first will import panders as PD beautiful and then we might create one from our

175
00:12:29,410 --> 00:12:30,880
existing array.

176
00:12:30,970 --> 00:12:35,890
So the data frame we're just gonna pass at A2.

177
00:12:35,960 --> 00:12:41,130
Let's have a quick look at what A2 is a A2 wonderful.

178
00:12:41,190 --> 00:12:42,390
So there's A2.

179
00:12:42,420 --> 00:12:43,940
Now let's see what that looks like.

180
00:12:44,010 --> 00:12:47,100
You need to view it shift and enter.

181
00:12:47,100 --> 00:12:48,300
There we go.

182
00:12:48,300 --> 00:12:52,470
So this is a panda's data frame and this is our num pi array.

183
00:12:52,470 --> 00:12:58,500
So essentially no matter what's in our Panda's data frame if it's all numbers and panders is built on

184
00:12:58,500 --> 00:13:06,930
top of num pie so to represent our data as a data frame it's just combining num higher res in a formatted

185
00:13:06,930 --> 00:13:08,070
way.

186
00:13:08,070 --> 00:13:13,020
So that's a very important point to remember as well is that once we start to get into the weeds of

187
00:13:13,020 --> 00:13:18,810
num pie and start to work out what kind of numerical computing it can do we can start to think okay.

188
00:13:18,930 --> 00:13:24,840
So if we have some data and we imported and panders it's actually just a different combination.

189
00:13:24,860 --> 00:13:30,900
Umpire res so if we want to perform some manipulations on it we're actually just performing manipulations

190
00:13:30,900 --> 00:13:37,380
on different lists of numbers horror as a brief introduction to num plays main data type.

191
00:13:37,380 --> 00:13:43,260
That's the main thing to take away from this is that NDA array whatever array you have in num pi is

192
00:13:43,260 --> 00:13:49,170
of type in the array and that panders and other machine learning algorithm packages are built on top

193
00:13:49,170 --> 00:13:55,130
of num pie and machine learning is finding patterns in num pi arrays.

194
00:13:55,170 --> 00:14:00,480
So with that being said let's take a little break if you need to go back and explore create an array

195
00:14:00,480 --> 00:14:02,360
of your own by trialling it out.

196
00:14:02,580 --> 00:14:05,250
If in doubt run the code and I'll see in the next video.