1
00:00:02,290 --> 00:00:10,450
So whenever we have the data in the tabular format ready with us, we should first look at the descriptions

2
00:00:10,450 --> 00:00:11,860
of each variable.

3
00:00:12,950 --> 00:00:19,550
The descriptive statistics, as we discussed in another section, are used to describe the data and

4
00:00:19,550 --> 00:00:20,270
summarize it.

5
00:00:21,490 --> 00:00:27,490
And since we will describe each and every single variable and not relationships between two or more

6
00:00:27,490 --> 00:00:34,050
variables, it is called univariate analysis unit stands for one very standard variable.

7
00:00:34,360 --> 00:00:36,820
So it is one variable analysis.

8
00:00:38,770 --> 00:00:43,120
So we can look at the bulletins as part of the univariate analysis.

9
00:00:44,380 --> 00:00:52,270
We can see mean median mode, we can see measures of dispersion like range quartiles and standard deviations.

10
00:00:53,760 --> 00:00:57,300
And for categorical data, we can look at count of each category.

11
00:00:58,850 --> 00:01:06,020
Most software packages for statistics have a very easy way to do univariate analysis for all the variables

12
00:01:06,020 --> 00:01:06,740
of the dataset.

13
00:01:07,990 --> 00:01:10,960
And when we run it, we see something like this.

14
00:01:15,370 --> 00:01:25,000
So for a variable like age, we can have info like mean median, minimum, maximum, the 25th fiftieth

15
00:01:25,000 --> 00:01:27,070
and 25th percentile values.

16
00:01:29,400 --> 00:01:37,500
Imagine if I arrange all the ages in ascending order, the first value will be the minimum value.

17
00:01:38,350 --> 00:01:44,250
The last value will be the maximum value, the 25th percentile value.

18
00:01:44,640 --> 00:01:49,520
This twenty five point seven five value will be at one fourth position.

19
00:01:50,040 --> 00:01:54,150
That is, 25 percent of the values will be lower than this value.

20
00:01:55,930 --> 00:02:00,440
Twenty nine days, 50 percent daily value will be at the middle.

21
00:02:00,700 --> 00:02:03,250
That is 50 percent of value will be lower than this.

22
00:02:04,000 --> 00:02:06,820
And this 50 percentile value is the same as the median.

23
00:02:06,820 --> 00:02:09,610
Also, if you remember the definition of median.

24
00:02:11,530 --> 00:02:20,980
Seventy five percent in value, 35, five to five is a Trefort position, so 75 percent values are below

25
00:02:20,980 --> 00:02:22,330
35 point to five.

26
00:02:23,560 --> 00:02:32,080
And from 75 percentile to the maximum value is thirty five point five to 51, this range is 75 percentile

27
00:02:32,080 --> 00:02:32,980
do maximum.

28
00:02:34,130 --> 00:02:36,620
So you can observe for this data.

29
00:02:37,780 --> 00:02:44,990
The first 25 percent values are in a very small range of 24 to twenty five point seven five.

30
00:02:46,450 --> 00:02:55,270
But if you see the last quartile, the last 25 percent of values are between a huge range of thirty

31
00:02:55,300 --> 00:02:57,370
five point two, five to fifty one.

32
00:02:58,330 --> 00:03:01,390
So this distribution is not evenly distributed.

33
00:03:03,340 --> 00:03:06,700
Such observations help us identify issues in data.

34
00:03:08,920 --> 00:03:16,210
This whole information for all the variables of the dataset is called the Extended Data Dictionary.

35
00:03:17,710 --> 00:03:25,630
Using day, we can use a lot of things like patterns of outliers, presence of missing values and so

36
00:03:25,630 --> 00:03:25,900
on.

37
00:03:27,420 --> 00:03:32,970
What are these issues and how we handle them will be covered in the coming videos?