1
00:00:01,430 --> 00:00:05,160
No, let us see how to do univariate analysis in our.

2
00:00:06,270 --> 00:00:11,190
We'll first find the extended data dictionary of each variable that is EDT.

3
00:00:11,520 --> 00:00:17,420
And after looking at EDT, if we have doubt on distribution of a particular valuable, we will plaudit

4
00:00:17,570 --> 00:00:20,440
its program to look at its distribution.

5
00:00:21,590 --> 00:00:24,740
Then we will also plot budgets of the categorical variables.

6
00:00:25,430 --> 00:00:28,910
And lastly, we will summarize our observations from the univariate analysis.

7
00:00:31,360 --> 00:00:32,390
Two good EDT.

8
00:00:32,450 --> 00:00:36,710
We just run somebody and within bracket relate D.F..

9
00:00:40,300 --> 00:00:41,160
And we done this.

10
00:00:41,430 --> 00:00:42,850
You can see in the window below.

11
00:00:43,320 --> 00:00:45,480
Let us make this window big.

12
00:00:45,870 --> 00:00:46,620
But looking here.

13
00:00:48,450 --> 00:00:58,750
So what each variable we are getting, the minimum, maximum mean and cortile values by quarter, I

14
00:00:58,750 --> 00:01:04,840
mean the first quarter, that is 25th percentile, the median, which is 58 percentile.

15
00:01:05,320 --> 00:01:07,270
And third quartile, which is 70 percent.

16
00:01:08,300 --> 00:01:15,340
So since this is a little bit difficult to read, because the same in value is going into two different

17
00:01:16,300 --> 00:01:18,820
roles for different types of variables.

18
00:01:19,690 --> 00:01:22,030
Let us put it into a readable format.

19
00:01:22,360 --> 00:01:24,940
We will reduce the font size for this.

20
00:01:25,610 --> 00:01:26,560
So go to view.

21
00:01:28,200 --> 00:01:34,570
Beans, and we'll go to being labeled with an appearance.

22
00:01:35,370 --> 00:01:37,830
We will reduce this font size from third to eleven.

23
00:01:39,210 --> 00:01:39,940
Click on to play.

24
00:01:42,010 --> 00:01:47,200
Let's click on Okina and we will extend this to the late.

25
00:01:48,320 --> 00:01:48,950
So that.

26
00:01:50,200 --> 00:01:51,880
It will die in order.

27
00:01:52,420 --> 00:01:52,690
So.

28
00:01:53,930 --> 00:01:54,650
Let's keep it here.

29
00:01:56,360 --> 00:01:57,920
Now it is visible clearly.

30
00:01:59,680 --> 00:02:04,490
As you know, the difference between mean and median values really indicate skewness and outliers.

31
00:02:05,320 --> 00:02:10,690
So if we look at price and it's mean and median values does seem fairly close.

32
00:02:12,120 --> 00:02:14,550
So we need not look at price.

33
00:02:15,590 --> 00:02:17,240
Crime rate, however, has.

34
00:02:18,780 --> 00:02:21,030
Bigger difference between mean and median values.

35
00:02:21,930 --> 00:02:26,130
Another indicator of Skewness and outliers is the distribution within quartiles.

36
00:02:26,730 --> 00:02:34,010
So if we look at the first quartile of crime rate within a short range of zero point zero six two zero

37
00:02:34,010 --> 00:02:39,420
point zero eight, in fact, did a one zero zero six two zero point zero eight.

38
00:02:40,540 --> 00:02:46,030
There is the first quartile that is, Gerti, five percent of values are within this short range.

39
00:02:47,220 --> 00:02:49,950
But the third quarter lendee, maximum value.

40
00:02:50,040 --> 00:02:51,450
That is the last quartile.

41
00:02:53,140 --> 00:02:56,410
It has a huge range from three point six to eighty eight point nine.

42
00:02:57,880 --> 00:02:59,800
And it also has 25 percent values.

43
00:03:00,520 --> 00:03:07,000
So there is a huge difference in the ranges and they both are containing only 25 percent of the values.

44
00:03:07,840 --> 00:03:13,210
So this very well either has outliers or skewness in its distribution.

45
00:03:14,680 --> 00:03:22,420
As we move ahead and look at each variable and its mean and median values, the next variable of interest

46
00:03:22,420 --> 00:03:23,860
is in hard rooms.

47
00:03:26,140 --> 00:03:30,190
And also the mean and within values have some difference.

48
00:03:30,370 --> 00:03:34,930
But if you look at the quartiles, the max value is exceptionally high.

49
00:03:35,320 --> 00:03:41,530
So from minimum of ten point zero 06 to third quartile, it is only going from 10 to 14.

50
00:03:42,190 --> 00:03:44,590
But from third quarter to maximum.

51
00:03:45,190 --> 00:03:47,350
It is moving from 14 to one zero one.

52
00:03:48,910 --> 00:03:53,200
So there is some distribution or outlier issue here.

53
00:03:54,450 --> 00:03:56,120
Similarly with rainfall.

54
00:03:57,670 --> 00:03:58,660
Minimum is three.

55
00:04:00,140 --> 00:04:02,800
And the first court date comes at twenty eight.

56
00:04:03,980 --> 00:04:11,340
But for all other quartiles, the distribution is nearly, what, a range of 10 unit to this minimum.

57
00:04:11,380 --> 00:04:15,560
Two, first cortile is a little bit skewed or has some outliers.

58
00:04:16,400 --> 00:04:24,370
So for crime rate and rooms and rainfall, we believe there is either Skewness in the distribution.

59
00:04:25,320 --> 00:04:26,670
Or presence of outliers.

60
00:04:27,560 --> 00:04:31,920
Not another thing to notice in this variable and horse, Biggs.

61
00:04:33,100 --> 00:04:34,650
It has one additional value of.

62
00:04:34,800 --> 00:04:37,190
And is there are eight any values?

63
00:04:38,110 --> 00:04:40,750
So while we import data from us, yes, we file.

64
00:04:40,870 --> 00:04:46,570
If there is any blank will, you are automatically assigned it a value of any.

65
00:04:48,070 --> 00:04:50,380
So we need to handle these missing values also.

66
00:04:51,220 --> 00:04:55,960
Now, let's get back to these three variables that we identified earlier, which either have outliers

67
00:04:55,990 --> 00:04:56,740
or Skewness.

68
00:04:57,520 --> 00:05:03,190
We need to see the actual distribution during the five which issued they are actually facing.

69
00:05:04,210 --> 00:05:07,930
The distribution can be seen in histograms or in scatterplot.

70
00:05:09,180 --> 00:05:11,490
First, we will plot histogram for each of these.

71
00:05:12,870 --> 00:05:13,940
To plot Instagram.

72
00:05:14,100 --> 00:05:15,600
Just write his.

73
00:05:17,930 --> 00:05:18,610
ATSDR.

74
00:05:20,280 --> 00:05:21,950
And and record, we relate.

75
00:05:26,050 --> 00:05:29,520
The F dollar crime under read.

76
00:05:36,910 --> 00:05:38,950
On the day, you can see the plot for this.

77
00:05:41,190 --> 00:05:47,250
If we look at the Instagram, it has bulk of values on the left between zero drippin probably.

78
00:05:48,360 --> 00:05:51,690
But how many values are going beyond the value of twenty?

79
00:05:51,810 --> 00:05:55,980
We are not really sure, even if there are 10 15 values.

80
00:05:56,370 --> 00:05:57,690
These may not be outliers.

81
00:05:57,720 --> 00:05:59,310
These may be genuine values.

82
00:05:59,940 --> 00:06:04,650
So this histogram is not giving us the correct picture for this particular variable.

83
00:06:05,130 --> 00:06:07,320
We should go and look at Scatterplot.

84
00:06:08,370 --> 00:06:16,650
So let's applaud Scatterplot for each of these three variables to get scatterplot for all of these simultaneously.

85
00:06:16,860 --> 00:06:21,180
We will rate payers and within bracket.

86
00:06:25,560 --> 00:06:31,990
We will start with DeLay and we'll start writing all the variables that we want in this bad.

87
00:06:32,390 --> 00:06:33,470
So first is price.

88
00:06:37,780 --> 00:06:38,290
Plus.

89
00:06:41,560 --> 00:06:45,970
Crime rate, close and heartbeat.

90
00:06:51,450 --> 00:06:53,850
Plus, rainfall, comma.

91
00:06:58,640 --> 00:07:00,820
All might be days, including if.

92
00:07:07,080 --> 00:07:07,500
Dundies.

93
00:07:15,180 --> 00:07:16,390
Let us correct this variable.

94
00:07:16,950 --> 00:07:18,060
It should be in hot rooms.

95
00:07:18,790 --> 00:07:19,650
Let's run it again.

96
00:07:21,120 --> 00:07:23,850
You can see we have a set of scatterplot here.

97
00:07:23,880 --> 00:07:26,160
Let us click on Zoom Button to look at them.

98
00:07:28,060 --> 00:07:28,890
That's Maximizer.

99
00:07:32,480 --> 00:07:36,200
So first, we need to look at price vs. crime rate.

100
00:07:37,350 --> 00:07:41,430
It has a lot of values within the range of zero to 10.

101
00:07:41,940 --> 00:07:50,280
But still a considerable number of values are going beyond this value took price and crime rate probably

102
00:07:50,280 --> 00:07:52,410
do not have a linear relationship.

103
00:07:52,980 --> 00:07:54,930
They have some other type of relationship.

104
00:07:57,380 --> 00:08:04,280
So we need to transform this variable in some way so that they end together leaner listenership.

105
00:08:04,640 --> 00:08:07,040
We will see this transformation a little reduce.

106
00:08:08,220 --> 00:08:11,160
Next, we will look at price and in hot rooms.

107
00:08:13,000 --> 00:08:20,650
If you look at this graph, all the values are within short range, but two values are exceptionally

108
00:08:20,740 --> 00:08:21,440
out of order.

109
00:08:23,190 --> 00:08:29,370
This exceptionally out of order values are certainly at outliers that do not behave in any particular

110
00:08:29,370 --> 00:08:29,850
pattern.

111
00:08:30,260 --> 00:08:31,770
They are clearly outliers.

112
00:08:32,250 --> 00:08:33,630
Similarly for rainfall.

113
00:08:35,840 --> 00:08:39,680
All of these values seem to be beyond 20 and within 60.

114
00:08:41,260 --> 00:08:44,260
Only one value is even below 10.

115
00:08:44,770 --> 00:08:46,750
This value is clearly an outlier.

116
00:08:47,560 --> 00:08:56,290
So we have identified that in hot rooms and rainfall has outliers crime rate as some different type

117
00:08:56,290 --> 00:08:57,580
of relationship with price.

118
00:08:57,760 --> 00:09:03,530
And we will manipulate this variable so that it has a linear type of relationship with price.

119
00:09:03,700 --> 00:09:04,210
Later on.

120
00:09:08,810 --> 00:09:12,950
One last thing we should do is looking at bar plots of categorical variables.

121
00:09:14,020 --> 00:09:15,790
We have three categorical variables.

122
00:09:16,820 --> 00:09:23,140
Airport, water, body and bus terminals, which is represented by bus and escort her.

123
00:09:25,040 --> 00:09:26,240
Do applaud, but applaud.

124
00:09:27,050 --> 00:09:27,750
We just right.

125
00:09:29,160 --> 00:09:29,990
But applaud.

126
00:09:31,490 --> 00:09:32,480
And within bracket.

127
00:09:37,390 --> 00:09:44,520
We'll right able and again within Blacket will specify the variable.

128
00:09:44,700 --> 00:09:47,230
It is D.F. Dollar Airport.

129
00:09:53,570 --> 00:09:54,350
Not Sundays.

130
00:09:56,930 --> 00:09:57,980
So you on the right.

131
00:09:58,160 --> 00:10:00,020
We have Bob blood of.

132
00:10:01,420 --> 00:10:04,270
Airport, which is a categorical variable as to values.

133
00:10:04,450 --> 00:10:05,380
Yes and no.

134
00:10:06,070 --> 00:10:12,190
And by looking at this graph, we do not see anything suspicious about this particular variable.

135
00:10:14,370 --> 00:10:16,560
No, let us do the same thing for everybody.

136
00:10:19,680 --> 00:10:20,640
Just change this very well.

137
00:10:20,670 --> 00:10:21,750
Do what, awardee?

138
00:10:26,160 --> 00:10:35,480
And on it, this also has forward values, Lake Lake, and we were bored none.

139
00:10:35,830 --> 00:10:36,400
And they were.

140
00:10:37,870 --> 00:10:39,690
And this also seems fine to us.

141
00:10:41,080 --> 00:10:44,130
Now, let us run Lebar platform bus terminal.

142
00:10:52,970 --> 00:11:00,660
Here you can see it has only one value on the cities offer of our dataset has bus terminal in them.

143
00:11:02,330 --> 00:11:09,490
So since all duties of our data have bus terminal, we cannot identify whether this variable impacts

144
00:11:09,490 --> 00:11:10,760
the final solution or not.

145
00:11:11,950 --> 00:11:14,740
So we need to ignore this variable in our analysis.

146
00:11:15,850 --> 00:11:22,930
So we did EDT, Blood-Red, Scatterplot and Bartletts to identify variables with outliers, skewness,

147
00:11:23,200 --> 00:11:25,840
missing values, and I use less categorical variable.

148
00:11:27,230 --> 00:11:28,840
So let's list Dondi observations.

149
00:11:30,610 --> 00:11:35,200
Posters and hard rooms and rainfall has outliers.

150
00:11:41,700 --> 00:11:45,390
And rainfall has always

151
00:11:48,780 --> 00:11:52,830
taken in hospitals as missing lose.

152
00:12:04,130 --> 00:12:08,240
It is bus terminal is a useless variable.

153
00:12:19,680 --> 00:12:22,220
And the last observation is crime rate.

154
00:12:25,930 --> 00:12:30,140
Crime rate has some other functional relationship with price.

155
00:12:40,350 --> 00:12:42,390
We'll handle these issues in the coming videos.