1
00:00:00,360 --> 00:00:07,350
Long before going deep dive into the session, let's have a quick recap of what we all have done in

2
00:00:07,350 --> 00:00:08,810
all our previous session.

3
00:00:09,150 --> 00:00:16,170
So from importing this data, we have the data from databases as well as from our seats flight that

4
00:00:16,170 --> 00:00:23,010
we have perform analysis on our data using this explo, which is exactly my sentiment analysis with

5
00:00:23,010 --> 00:00:31,350
respect to this summary feature after we have analyzed here what exactly the expatriated analysis for

6
00:00:31,680 --> 00:00:34,870
positive sentiment or you can say positive sentences.

7
00:00:34,890 --> 00:00:40,470
So this is exactly the accelerated analysis for our positive sentiment.

8
00:00:40,530 --> 00:00:42,690
So in this session, we have this assignment.

9
00:00:42,690 --> 00:00:49,650
The very first one is we have to perform this exploratory analysis for the negative sentences.

10
00:00:49,650 --> 00:00:57,570
And this is exactly the second statement in which I have to analyze to what type of users Amazon can

11
00:00:57,570 --> 00:00:59,540
recommend more product.

12
00:00:59,670 --> 00:01:02,430
So it's a very popular use case of Amazon.

13
00:01:02,430 --> 00:01:04,840
And believe me, it's a very popular use case of Amazon.

14
00:01:05,310 --> 00:01:11,150
So so, yeah, let's let's go ahead with very first statement in which you have to perform this analysis

15
00:01:11,700 --> 00:01:17,370
so what you guys can do rather than writing code again and again, I'm going to share your trick.

16
00:01:17,370 --> 00:01:19,340
Just copy and paste.

17
00:01:19,710 --> 00:01:20,090
Yeah.

18
00:01:20,100 --> 00:01:23,870
Because most of the time you have to do this as well.

19
00:01:23,880 --> 00:01:29,370
So I'm just going to copy and just going to paste, just do some modifications.

20
00:01:29,730 --> 00:01:33,510
That said, here I am going to say it is less than zero.

21
00:01:33,510 --> 00:01:40,740
And here I will say this time my data is nothing but data on this go negative.

22
00:01:40,770 --> 00:01:48,660
So now what we have to do in this huge chunk of data, you will figure out you have to just combine

23
00:01:48,660 --> 00:01:50,880
this huge chunk of data here.

24
00:01:50,880 --> 00:01:58,350
I'm going to say I have to combine now, but make sure you have to combine in this data on a score negative

25
00:01:58,350 --> 00:01:58,630
data.

26
00:01:58,920 --> 00:02:01,560
So I have to just combine all this stuff here.

27
00:02:01,560 --> 00:02:06,150
I'm going to say this my total and a score text, too, so just executed.

28
00:02:06,180 --> 00:02:10,500
Now what you have to do, you have to perform some cleaning on your data.

29
00:02:10,650 --> 00:02:13,320
So which is exactly this one.

30
00:02:13,560 --> 00:02:21,140
So now I have to just paste over here and this time I'm going to save my data is total tax too.

31
00:02:21,420 --> 00:02:25,110
And this is also total tax to now what we have to do.

32
00:02:25,110 --> 00:02:29,850
We have to remove some extra spaces if it is available in data.

33
00:02:30,030 --> 00:02:33,530
So let me check if it is available in data.

34
00:02:33,560 --> 00:02:34,260
Not so.

35
00:02:34,260 --> 00:02:37,090
You will figure out you still have some extra spaces.

36
00:02:37,380 --> 00:02:44,700
So now what you have to do, you have to just copy these blocks of code from here and you have to just

37
00:02:44,910 --> 00:02:46,650
paste over here this time.

38
00:02:46,650 --> 00:02:53,770
I would say it is my tool and this is also my total in this context on a scale to just executed.

39
00:02:53,790 --> 00:02:59,990
Now it's time for to generate your beautiful visual, which is exactly my workout.

40
00:03:00,270 --> 00:03:06,960
So here I have to say my children to school text to this time this is my word cloud to this in my word

41
00:03:06,960 --> 00:03:08,130
cloud to here.

42
00:03:08,130 --> 00:03:14,460
I had to slide my word cloud two and after it all is of similar just executed.

43
00:03:14,460 --> 00:03:20,400
It will take some couple of seconds to showcase that beautiful word out in front of you.

44
00:03:20,430 --> 00:03:23,250
So it is exactly that word cloud here.

45
00:03:23,250 --> 00:03:26,430
You will see the negative sentiment customers.

46
00:03:26,580 --> 00:03:34,830
They are basically going to use these words, say disappointed, bad taste, horrible, terrible, expansive,

47
00:03:34,830 --> 00:03:44,150
expected flavor, little and all these kind of keywords are user or customer are going to Praful.

48
00:03:44,310 --> 00:03:50,970
So now let's move to an exponent statement in which you have to perform analysis and you have to come

49
00:03:50,970 --> 00:03:56,970
up with some conclusion to what type of user Amazon recommend more products.

50
00:03:57,240 --> 00:03:59,520
So far, this is what I'm going to do.

51
00:03:59,550 --> 00:04:07,860
So very first you have to understand to who or you can see what type of users Amazon recommend products.

52
00:04:07,860 --> 00:04:15,450
So Amazon blackmond more products to only those one who are going to buy more or two, those one who

53
00:04:15,450 --> 00:04:17,100
has a better conversion rate.

54
00:04:17,430 --> 00:04:22,380
So very first to have to make your data ready for your analysis purpose.

55
00:04:22,740 --> 00:04:29,550
So what I am going to do in this, let me show you are things if I'm going to call this dot head or

56
00:04:29,550 --> 00:04:31,860
you can also call did it or it's all up to you.

57
00:04:32,310 --> 00:04:39,570
So in this doctor had you will see here you have a column, this user I'd seen this user idea.

58
00:04:39,960 --> 00:04:45,920
You have to search those users to which Amazon is going to recommend product.

59
00:04:46,170 --> 00:04:54,630
So in this user I.D., if I have to ask this user ID and on this, if I am going to call my end unique

60
00:04:54,750 --> 00:04:59,970
number of unique user, you will observe it has that much unique.

61
00:05:00,450 --> 00:05:06,960
So from that huge chunk of yuzu, you have to extract some top end users, you have to extract some

62
00:05:07,260 --> 00:05:10,940
top 20 users depending upon different different qualities.

63
00:05:10,950 --> 00:05:13,500
They have different different specifications.

64
00:05:13,500 --> 00:05:15,710
They have like more number of products.

65
00:05:15,720 --> 00:05:18,420
Those users are going to buy more.

66
00:05:18,450 --> 00:05:24,540
The recommendation that a user have for it means you have to up this huge chunk of data on the basis

67
00:05:24,540 --> 00:05:33,300
of this user I.D. So here I am going to say, does dot group buy on the basis of this user ID?

68
00:05:33,300 --> 00:05:36,290
I have to group my data once I recoup my data.

69
00:05:36,510 --> 00:05:39,200
I had to aggregate my all of the features.

70
00:05:39,390 --> 00:05:46,530
So here I am going to say, oh, I need total summary of each and every user.

71
00:05:46,560 --> 00:05:49,860
I need this total summary of what I can say.

72
00:05:49,860 --> 00:05:54,820
I need count of summary of each and every user.

73
00:05:54,840 --> 00:06:03,090
So here I am going to say I have to aggregate and I have to aggregate the summary as by performing my

74
00:06:03,120 --> 00:06:05,360
count toward you in a similar way.

75
00:06:05,400 --> 00:06:15,240
I'm going to say on this text I'm also going to perform my count because I need every user has given

76
00:06:15,240 --> 00:06:17,430
how much number of feedbacks.

77
00:06:17,640 --> 00:06:24,300
So I'm going to say on this text, I have to just perform this count operation.

78
00:06:24,600 --> 00:06:31,890
And after it, I'm just going to say I just need a meniscal of each and every user.

79
00:06:31,900 --> 00:06:36,410
So here I am going to say I need a mean score.

80
00:06:36,420 --> 00:06:37,580
So is goal.

81
00:06:37,890 --> 00:06:42,050
And here you have to mention that data in the form of dictionary.

82
00:06:42,330 --> 00:06:50,970
And now what you have to do in this product ID, you have to perform again, gone because I have to

83
00:06:50,970 --> 00:06:53,940
find you each and every user has.

84
00:06:53,940 --> 00:06:56,370
But how many number of products?

85
00:06:56,700 --> 00:07:05,010
So here I am going to say on this product ID, I have to just perform this count operation now if I'm

86
00:07:05,010 --> 00:07:07,140
going to execute this blocks of code.

87
00:07:07,150 --> 00:07:14,080
So it will exactly return me this beautiful statistics on which you have to analyze this data.

88
00:07:14,430 --> 00:07:17,430
So let's say I have to short this data frame.

89
00:07:17,430 --> 00:07:19,580
Let's have to short this all the stuff.

90
00:07:19,860 --> 00:07:24,570
So for this, I'm going to say not short on this core values.

91
00:07:24,570 --> 00:07:29,090
And here you have a parameter which is exactly your by.

92
00:07:29,310 --> 00:07:35,820
So here I am going to say I have to short this on the basis of basically this text.

93
00:07:36,240 --> 00:07:42,090
I'm going to say I showed this on the basis of X, then I have to show this data any form of, let's

94
00:07:42,090 --> 00:07:43,770
say, descending order for this.

95
00:07:43,770 --> 00:07:51,480
I have to set my ascending parameter as false because by default it decides to just execute it again.

96
00:07:51,480 --> 00:07:53,620
It will take some couple of seconds.

97
00:07:53,640 --> 00:08:00,660
Now, in this table, you will figure out this user has that much number of somebody that has given

98
00:08:00,660 --> 00:08:02,160
that much number of feedback's.

99
00:08:02,940 --> 00:08:09,860
This user has that much meniscal and he has bought that much number of product.

100
00:08:10,080 --> 00:08:14,650
So it is a beautiful statistics with respect to each and every user I.D..

101
00:08:15,000 --> 00:08:18,510
Now, you will say that all these are exactly my top 10 users.

102
00:08:18,510 --> 00:08:21,630
You will see let's say I have to store it somewhere.

103
00:08:21,750 --> 00:08:24,960
Let's say I will say it is my let's say recorded of him.

104
00:08:24,960 --> 00:08:25,330
It's all up.

105
00:08:25,440 --> 00:08:30,990
You know what I'm going to do if I'm going to print now in this raw data frame?

106
00:08:31,200 --> 00:08:33,280
What we have to do, we have to manage.

107
00:08:33,510 --> 00:08:35,180
It is still the column names.

108
00:08:35,490 --> 00:08:39,480
So here I am going to say are the DOT columns.

109
00:08:39,480 --> 00:08:43,120
Let's say I have to assign my own column name next to the very first column.

110
00:08:43,120 --> 00:08:49,770
Name that I have to assign is exactly my, let's say, number of some B so here I'm going to say number

111
00:08:49,770 --> 00:08:52,830
of summaries.

112
00:08:53,010 --> 00:09:01,200
The second column name that I want to assign is exactly my let's say num underscore text or number of

113
00:09:01,350 --> 00:09:01,830
text.

114
00:09:02,310 --> 00:09:10,830
The third one is exactly my let's say average on this score, average score or you can say it is meniscal

115
00:09:10,830 --> 00:09:13,520
and the fourth column, which is exactly as it is.

116
00:09:13,720 --> 00:09:23,550
I have to assign, let's say, number of product that are usable, just our number of products user

117
00:09:23,550 --> 00:09:25,380
purchased are purchased.

118
00:09:25,380 --> 00:09:28,280
It's all appeal whatever column name you want to assign.

119
00:09:28,680 --> 00:09:34,790
Now, after doing all these things, let's say I have to just print my data and just execute it.

120
00:09:34,800 --> 00:09:40,640
This is that finally the frame that you have to consider for your visual purpose.

121
00:09:40,650 --> 00:09:47,190
Now, let's say from this huge chunk of data, I just need my top ten users.

122
00:09:47,430 --> 00:09:53,910
So I'm just going to say I'm going to use my body plot from my matplotlib.

123
00:09:53,910 --> 00:09:59,430
So I'm going to save BLT Dot Dot and here, if you will, press crosstab.

124
00:09:59,720 --> 00:10:06,200
You will figure out all these custom parameters of the sanction, what is X, what is all these kinds

125
00:10:06,200 --> 00:10:06,520
of things?

126
00:10:06,560 --> 00:10:13,340
What is X height, weight and all these things, let's say on X axis?

127
00:10:13,580 --> 00:10:18,050
I just need the user I.D. for this.

128
00:10:18,050 --> 00:10:24,470
What I'm going to do, I'm just going to say it is nothing but my, let's say, raw dot index.

129
00:10:24,470 --> 00:10:28,160
And in this index I just need my top 10 index.

130
00:10:28,380 --> 00:10:30,350
Let's say I have to store it somewhere.

131
00:10:30,740 --> 00:10:34,970
So for design, I'm say to nothing but my user and the score then.

132
00:10:35,270 --> 00:10:36,410
So just executed.

133
00:10:36,440 --> 00:10:43,610
Now what we have to do very first on this X, we have to assign this user 10 and now we have some other

134
00:10:43,610 --> 00:10:44,410
parameter.

135
00:10:44,450 --> 00:10:48,080
The second parameter, the most important one is exactly your height.

136
00:10:48,410 --> 00:10:52,040
Let's say in this height, I just need this count.

137
00:10:52,310 --> 00:10:56,980
So here I'm going to say it is nothing but my raw all.

138
00:10:57,170 --> 00:11:02,950
Let's say a number of or you can say number of product purchased.

139
00:11:03,200 --> 00:11:11,660
So here I will say it is nothing but a number of punches and I just need to top ten people, whatever

140
00:11:11,990 --> 00:11:14,870
you want, I have to store it somewhere else.

141
00:11:14,870 --> 00:11:16,280
So I'm going to say it is nothing.

142
00:11:16,280 --> 00:11:20,520
But my number underscored ten just executed.

143
00:11:20,540 --> 00:11:24,880
Now you have to pass this number underscored that over here.

144
00:11:25,280 --> 00:11:28,130
Now you have some additional parameters.

145
00:11:28,130 --> 00:11:34,550
Let's say what label you want to assign to my label is nothing but my let's say whatever label you want

146
00:11:34,550 --> 00:11:37,400
to assign, let's say most recommended users.

147
00:11:37,400 --> 00:11:42,560
So I'm going to say most are recommended users.

148
00:11:42,560 --> 00:11:46,120
And after it, what I am going to do, I'm just going to execute it.

149
00:11:46,120 --> 00:11:47,750
It will take some couple of seconds.

150
00:11:48,020 --> 00:11:55,190
You will still see over here this graph looks a little bit messy and still you don't have some extra

151
00:11:55,190 --> 00:11:56,090
button available.

152
00:11:56,360 --> 00:12:03,950
So for this, let's say I'm just going to assign some X level using this BLT dot X label.

153
00:12:04,190 --> 00:12:12,290
And let's say my ex label is, let's say user I.D. And after it, we have to assign some white label

154
00:12:12,290 --> 00:12:12,830
as well.

155
00:12:12,830 --> 00:12:17,570
So my label is nothing but let's say no product Porchester here.

156
00:12:17,570 --> 00:12:26,290
I would say number of dollars for change and after that we have to rotate our X label as well.

157
00:12:26,540 --> 00:12:35,890
So for this I would say BLT dot, x, x and here I have a barometer which is exactly my rotation parameters.

158
00:12:35,910 --> 00:12:38,450
Here I'm going to say rotation.

159
00:12:38,720 --> 00:12:41,300
It goes to all the girls.

160
00:12:41,690 --> 00:12:43,460
So execute it.

161
00:12:43,460 --> 00:12:45,070
It will take some couple of seconds.

162
00:12:45,440 --> 00:12:49,430
And this is a beautiful stat that you exactly need.

163
00:12:49,730 --> 00:12:57,200
And if you want some more beautiful with additions, you guys can go ahead with module that we all are

164
00:12:57,200 --> 00:13:00,950
going to cover in all our upcoming projects.

165
00:13:00,950 --> 00:13:09,590
So from this board plot, you will conclude these are exactly my top ten users so we can recommend more

166
00:13:09,590 --> 00:13:18,020
and more product to this user ID as there is a high probability that this person are going to buy the

167
00:13:18,020 --> 00:13:18,620
product.

168
00:13:19,190 --> 00:13:21,320
So that's all about this conclusion.

169
00:13:21,320 --> 00:13:23,940
That's all about this problem statement as well.

170
00:13:23,940 --> 00:13:26,090
I hope you love the session very much.

171
00:13:26,390 --> 00:13:27,080
Thank you.

172
00:13:27,080 --> 00:13:28,060
Have a nice day.

173
00:13:28,250 --> 00:13:29,090
Keep learning.

174
00:13:29,090 --> 00:13:31,160
Keep growing, keep motivating.