1
00:00:00,330 --> 00:00:06,270
All in all, our position, we have gone through with lots of analysis from importing the data from

2
00:00:06,270 --> 00:00:12,330
database, and at times we have performed some sentiment analysis with respect to positive, with respect

3
00:00:12,330 --> 00:00:16,950
to negative lots of things, this word cloud and not something.

4
00:00:17,220 --> 00:00:23,670
And this is your what are my top 10 users to which Amazon is going to recommend product.

5
00:00:24,230 --> 00:00:32,580
So session the very first part of the statement is we have to analyze the length of the comment, whatever

6
00:00:32,580 --> 00:00:34,210
the user are going to give.

7
00:00:34,470 --> 00:00:38,420
So we have to analyze whether they are lendee over that short one.

8
00:00:38,430 --> 00:00:42,110
Or you can say you can analyze its distribution as well.

9
00:00:42,720 --> 00:00:51,990
So maybe some of you have not that much good specification in terms of the processor or hard disk.

10
00:00:52,260 --> 00:00:59,220
So you're according to your system specifications, you can consider some sample of this huge chunk

11
00:00:59,220 --> 00:01:04,640
of data because you will figure out this is very huge data so you can consider some sample of data.

12
00:01:04,650 --> 00:01:10,100
And yeah, if you don't have some issue with the specification, just go with this bulky data.

13
00:01:10,110 --> 00:01:11,340
It's all up to you literally.

14
00:01:11,340 --> 00:01:16,670
It's all about you say for your sake, I'm going to consider some sample of data.

15
00:01:16,980 --> 00:01:23,310
So here I'm going to say either you can consider some sample of data using this data sample what our

16
00:01:23,310 --> 00:01:30,510
sample size you want to assign using this, you can get some sample or you guys can consider some sample

17
00:01:30,510 --> 00:01:31,280
as legacy.

18
00:01:31,350 --> 00:01:36,170
And when you consider some top two thousand rows, it's all up to you.

19
00:01:36,990 --> 00:01:42,630
Let's say this is my final little frame, just executed now is on this final.

20
00:01:42,630 --> 00:01:43,720
I'm going to call ahead.

21
00:01:44,040 --> 00:01:47,610
So this is that data frame on which I had to perform this analysis.

22
00:01:48,370 --> 00:01:51,750
Check whether I have any missing value in this data or not for this.

23
00:01:51,750 --> 00:02:00,290
I'm just trying to say is and a dot some are you guys can also use this is not awesome.

24
00:02:00,300 --> 00:02:01,170
It's all up to you.

25
00:02:01,470 --> 00:02:04,680
And on this, I'm just going to call this Willetton.

26
00:02:04,680 --> 00:02:08,070
Maiya, I don't have any missing value in my data.

27
00:02:08,200 --> 00:02:13,530
Now, let me check whether if I have any duplicates in or not.

28
00:02:13,530 --> 00:02:18,560
So for this, I'm just going to say final dot.

29
00:02:18,570 --> 00:02:24,750
I have a function which is exactly duplicated, dark, some just executed.

30
00:02:24,780 --> 00:02:26,110
Now it will return me.

31
00:02:26,250 --> 00:02:26,700
Yeah.

32
00:02:26,700 --> 00:02:29,590
You don't have any duplicates in your data.

33
00:02:29,880 --> 00:02:35,130
Now what you have to do, you have to analyze the land of the customers comment.

34
00:02:35,370 --> 00:02:42,900
So now what I am going to do, I'm just going to say in this text column, I have to extract what exactly

35
00:02:42,900 --> 00:02:47,230
the length of each of the feedback given by my customer.

36
00:02:47,430 --> 00:02:52,070
So here I am going to say, let's say final of text.

37
00:02:52,080 --> 00:02:52,680
Very first.

38
00:02:52,680 --> 00:02:53,820
I have to access this.

39
00:02:54,120 --> 00:02:57,180
Let's say I'm just going to access my zeroth index data.

40
00:02:57,180 --> 00:02:57,960
Just paint it.

41
00:02:57,960 --> 00:03:03,290
You will see it has that much number of words in this feedback.

42
00:03:03,540 --> 00:03:05,550
Let's say I have to split it.

43
00:03:05,550 --> 00:03:10,530
Let's say I have to split it on the basis of, let's say, space separated, just executed.

44
00:03:10,530 --> 00:03:12,690
It will be done with this list.

45
00:03:12,930 --> 00:03:18,850
And if on this list I'm going to calculate the length of this entire list.

46
00:03:18,990 --> 00:03:24,510
Now, you will figure out this combined has that much land in a similar way.

47
00:03:24,510 --> 00:03:28,590
You can compute the length of each and every comment.

48
00:03:29,040 --> 00:03:35,670
So for this, you guys can define a function to do this operation in a much more user friendly way.

49
00:03:35,700 --> 00:03:42,660
So here I'm going to say calculate land and whatever text I'm going to pass to my function, it will

50
00:03:42,660 --> 00:03:44,810
exactly return me my land.

51
00:03:45,090 --> 00:03:47,880
So here I am going to say whatever text I have on this.

52
00:03:47,880 --> 00:03:52,680
Very first I have to close on the split and I have to split it on the basis of the separator.

53
00:03:53,010 --> 00:03:59,880
Once I have all this stuff that I'm going to compute land the simple that just like a piece of OK,

54
00:03:59,890 --> 00:04:03,370
now whatever land you have, you have to just return it.

55
00:04:03,390 --> 00:04:04,190
That's it.

56
00:04:04,230 --> 00:04:05,880
That's a task of the function.

57
00:04:06,180 --> 00:04:13,320
Now what you have do you have to map dysfunction or you can say you have to apply this function on your

58
00:04:13,560 --> 00:04:14,580
text column.

59
00:04:14,590 --> 00:04:23,250
So final text dot apply and here you have to say calc underscore, just press tab.

60
00:04:23,710 --> 00:04:25,850
So it is exactly calcaneus colon.

61
00:04:26,280 --> 00:04:30,270
Now whatever land it will hit on me, I'm going to store it somewhere else.

62
00:04:30,270 --> 00:04:32,160
Let's say I define a column for that.

63
00:04:32,490 --> 00:04:39,170
So I'm going to say it is nothing but text, underscore and whatever name you want to assign, just

64
00:04:39,180 --> 00:04:42,030
execute it and it will give me some warning.

65
00:04:42,030 --> 00:04:43,680
Just don't worry at all.

66
00:04:43,680 --> 00:04:46,500
It is just because of some your anaconda's choose.

67
00:04:46,800 --> 00:04:49,980
Some of you have some personal issues, library issues.

68
00:04:50,220 --> 00:04:52,440
It is just because of so many issues.

69
00:04:52,440 --> 00:04:53,400
Just ignore this.

70
00:04:53,400 --> 00:04:56,430
Wanting to know what you have to do.

71
00:04:56,430 --> 00:04:59,790
You have to basically let's say I just need.

72
00:05:00,160 --> 00:05:08,510
Distribution of this tax land so far, this you guys can use a very handy function, which is exactly

73
00:05:08,510 --> 00:05:16,220
my box plot, so very close I have to import my Lochley because I want to use that box dot, which is

74
00:05:16,220 --> 00:05:18,070
exactly in my plot module.

75
00:05:18,080 --> 00:05:27,070
You can also use box stored in your Poinar Seabourne and many other things, but I just need some user

76
00:05:27,230 --> 00:05:28,190
friendly visuals.

77
00:05:28,220 --> 00:05:30,180
That's one way to use it here.

78
00:05:30,200 --> 00:05:39,050
So I'm going to say Pratley Dot Xpress as a bit and if you haven't told us you can install using this

79
00:05:39,320 --> 00:05:47,600
PIP install, pip install what the US can install using this basic command.

80
00:05:47,960 --> 00:05:50,780
Now what I have to do, I have two very first imported.

81
00:05:51,020 --> 00:05:56,930
Now using this I have a function which is exactly box just passive gustad.

82
00:05:56,930 --> 00:05:59,850
You have all these different different parameters Saussure.

83
00:06:00,240 --> 00:06:04,250
Now the very first parameter is exactly what is the frame which is final.

84
00:06:04,760 --> 00:06:14,480
And on this Y-axis, I'm just going to say I have to pass this next underscore land just executed.

85
00:06:14,480 --> 00:06:21,120
It will take some couple of seconds and it will return as my beautiful box brought in a while.

86
00:06:21,200 --> 00:06:25,050
So this is exactly that box, part of which I'm talking about.

87
00:06:25,340 --> 00:06:30,020
So if you have to conclude from this box plot, you will fill it out.

88
00:06:30,410 --> 00:06:37,790
It seems to have almost 50 percent users are going to give their feedback almost almost.

89
00:06:37,790 --> 00:06:43,250
You will figure it out, almost 50 words where there are only few users.

90
00:06:43,430 --> 00:06:50,240
You will see there are a few users who are going to give a word Landi feedback's, because these DOTD

91
00:06:50,660 --> 00:06:53,590
are exactly my outliers over here.

92
00:06:54,140 --> 00:06:58,640
So that's a conclusion that I have derived from here, from my data.

93
00:06:59,000 --> 00:07:04,850
So let's go ahead with our next thought on a statement in which I have to analyze our score.

94
00:07:05,150 --> 00:07:09,290
So in this final data frame here, you will figure it out.

95
00:07:09,860 --> 00:07:12,320
Here you have a column name of the score.

96
00:07:12,680 --> 00:07:16,160
So you have to analyze this feature for this.

97
00:07:16,170 --> 00:07:20,480
What I'm going to do, I'm just going to call my control over here.

98
00:07:20,480 --> 00:07:28,670
So as dot com plot and here I'm just going to say final score just executed.

99
00:07:28,670 --> 00:07:30,560
It will return this beautiful plot.

100
00:07:31,010 --> 00:07:38,360
So from this, you guys can figure out most of the customers are going to give us five score on any

101
00:07:38,360 --> 00:07:39,160
of the product.

102
00:07:39,560 --> 00:07:44,570
So that's the type of conclusion you can fetch from all the visas that you have over here.

103
00:07:44,930 --> 00:07:53,390
But believe me, 70 to 80 percent of the time is still gas is stand in your data preprocessing and in

104
00:07:53,390 --> 00:07:53,840
your data.

105
00:07:54,230 --> 00:08:00,440
That's what I had learned from my own experience working in analytics domain and in data science industry.

106
00:08:01,040 --> 00:08:03,350
So I hope you will love this session very much.

107
00:08:03,680 --> 00:08:04,440
Thank you.

108
00:08:04,490 --> 00:08:05,510
Have a nice day.

109
00:08:05,540 --> 00:08:06,380
Keep learning.

110
00:08:06,380 --> 00:08:07,190
Keep growing.

111
00:08:07,610 --> 00:08:08,450
Keep practicing.