1
00:00:00,066 --> 00:00:00,999


2
00:00:01,000 --> 00:00:07,700
Hello and welcome to a video on histograms. We want to use a histogram when we'd like to see the shape of our distribution.

3
00:00:07,700 --> 00:00:08,100


4
00:00:08,100 --> 00:00:14,100
Histogram show frequency data, which means they show how many of a certain type of data we have in set ranges called bins.

5
00:00:14,100 --> 00:00:14,566


6
00:00:14,566 --> 00:00:20,966
Let's start by loading in some data. So first we'll load our libraries. We have pandas as PD and from Google.colab 

7
00:00:20,966 --> 00:00:20,999


8
00:00:21,000 --> 00:00:27,166
we're importing drive, and these are the two that we'll use to actually set our file path in import

9
00:00:27,166 --> 00:00:33,332
our Excel file as a data frame. I'm using NumPy as NP to generate some

10
00:00:33,333 --> 00:00:33,366


11
00:00:33,366 --> 00:00:39,199
random numbers below in an example histogram, but you won't use that in your homework more than likely for your

12
00:00:39,200 --> 00:00:39,433


13
00:00:39,433 --> 00:00:45,633
histogram at least. We're also importing Matplotlib.pyplot as PLT and this is really

14
00:00:45,633 --> 00:00:45,999


15
00:00:46,000 --> 00:00:52,133
important because it's what we're going to use to generate our histogram. Let's go ahead

16
00:00:52,133 --> 00:00:52,166


17
00:00:52,166 --> 00:00:58,232
and run this block of code and get our variables all set in place. Now we have IV1, IV2 and DV available

18
00:00:58,233 --> 00:01:04,233
to us. Now what if we would like to make a histogram? Let's see what one looks like

19
00:01:04,233 --> 00:01:10,433
with some simple example data. Then we'll make a histogram for IV1. Remember histogram show

20
00:01:10,433 --> 00:01:16,599
frequency data. So this section here is just to generate some random numbers. Let's

21
00:01:16,600 --> 00:01:22,866
actually set this equal to 1,000. Next

22
00:01:22,866 --> 00:01:22,899


23
00:01:22,900 --> 00:01:28,800
we have PLT.Hist and that is the function that you'll use to create histograms.

24
00:01:28,800 --> 00:01:30,433


25
00:01:30,433 --> 00:01:36,433
We'll put data where data is supposed to go. Set our number of bins equal to 30. Set the color to sky

26
00:01:36,433 --> 00:01:36,466


27
00:01:36,466 --> 00:01:41,199
skyblue and the edge color to black. Next we'll add titles and labels. We're deciding

28
00:01:41,200 --> 00:01:43,300


29
00:01:43,300 --> 00:01:49,500
to show the grid with this plot but we could set this to false if we wanted to. And we'll display the plot with PLT.show.

30
00:01:49,500 --> 00:01:49,533


31
00:01:49,533 --> 00:01:55,466
Let's go ahead and run this code block. And you can see now we have

32
00:01:55,466 --> 00:01:55,732


33
00:01:55,733 --> 00:02:01,933
a histogram of random data. What's interesting about generating random numbers

34
00:02:01,933 --> 00:02:07,966
like this is that as you generate more and more numbers, your histogram will become more and more normally

35
00:02:07,966 --> 00:02:10,866
distributed. So if I were to set this to 250,

36
00:02:10,866 --> 00:02:16,499


37
00:02:16,500 --> 00:02:19,100
you can see that it's a little less normally distributed.

38
00:02:19,100 --> 00:02:22,533


39
00:02:22,533 --> 00:02:25,366
If I were to set this to 100, we can see

40
00:02:25,366 --> 00:02:28,899


41
00:02:28,900 --> 00:02:34,933
it's even less normally distributed. If we were to set this to 2000,

42
00:02:34,933 --> 00:02:38,999


43
00:02:39,000 --> 00:02:44,366
now we can see it's very normally distributed. And that is just the case when you generate random numbers.

44
00:02:44,366 --> 00:02:45,366


45
00:02:45,366 --> 00:02:50,799
This is part of something called the central limit theorem, which is something that you'll study very briefly in class.

46
00:02:50,800 --> 00:02:51,833


47
00:02:51,833 --> 00:02:58,133
But it's really cool to see an actual visual example of it working. So if you're curious

48
00:02:58,133 --> 00:02:58,166


49
00:02:58,166 --> 00:03:04,532
about whether you can use a color other in the sky blue, you can. You can check out the colors available on this website

50
00:03:04,533 --> 00:03:04,633


51
00:03:04,633 --> 00:03:10,866
or you can just Google what colors can I use in Python and it will come up with several different websites that will

52
00:03:10,866 --> 00:03:10,899


53
00:03:10,900 --> 00:03:17,100
show you examples of colors you can use. I just like this one because it lists both

54
00:03:17,100 --> 00:03:17,133


55
00:03:17,133 --> 00:03:22,999
the simple colors you can use and also the more involved colors you can do things like coral

56
00:03:23,000 --> 00:03:23,133


57
00:03:23,133 --> 00:03:28,499
or chocolate or green yellow or whatever speaks to you.

58
00:03:28,500 --> 00:03:32,233


59
00:03:32,233 --> 00:03:38,266
Okay now let's take a look at some real data using IV1. So we're going to set data equal to df IV1 this

60
00:03:38,266 --> 00:03:44,099
time. You don't have to do it this way. We've got data right here. Which is being

61
00:03:44,100 --> 00:03:44,533


62
00:03:44,533 --> 00:03:50,533
called by data above. We could actually just set this equal to df['IV1'] right here. But for

63
00:03:50,533 --> 00:03:51,733


64
00:03:51,733 --> 00:03:57,899
the sake of consistency with the code above we're going to set that equal to data.

65
00:03:57,900 --> 00:04:00,933


66
00:04:00,933 --> 00:04:07,033
the bins equal to 30. The color is sky blue and the edge color is black. Next we'll set

67
00:04:07,033 --> 00:04:07,066


68
00:04:07,066 --> 00:04:12,566
our title, x-label, y-label, grid equal to true, and display the plot.

69
00:04:12,566 --> 00:04:13,266


70
00:04:13,266 --> 00:04:19,299
Let's go. Alright, that's pretty different. We can see in the first example

71
00:04:19,300 --> 00:04:25,466
for a histogram that the data is very normally distributed, whereas in this second example it's more positively skewed. We've

72
00:04:25,466 --> 00:04:31,532
got some skewness to our data this time. Interesting. This tells us something about our variable that we did not

73
00:04:31,533 --> 00:04:31,566


74
00:04:31,566 --> 00:04:37,599
know previously and may affect how we perform statistical analysis on it. It's always

75
00:04:37,600 --> 00:04:43,700
important to visualize your data before you actually do any kind of analysis. Now, what if we want our histogram to look

76
00:04:43,700 --> 00:04:47,933
a little different? We could set our bins equal to 20 and

77
00:04:47,933 --> 00:04:49,866


78
00:04:49,866 --> 00:04:56,099
see how that affects the overall shape. I think I preferred

79
00:04:56,100 --> 00:04:56,133


80
00:04:56,133 --> 00:04:59,499
30 because we

81
00:04:59,500 --> 00:05:07,733


82
00:05:07,733 --> 00:05:13,733
can see that we have a drop off here now. But that's just a matter of personal preference.

83
00:05:13,733 --> 00:05:14,233


84
00:05:14,233 --> 00:05:20,466
So long as you have more than 10 or so bins, and you have fewer bins than you actually

85
00:05:20,466 --> 00:05:26,499
have in your distribution. So for example, we have a 170 data points in our data

86
00:05:26,500 --> 00:05:32,400
frame that we have loaded in. We don't want 170 bins. I'll show you what that looks like

87
00:05:32,400 --> 00:05:32,933


88
00:05:32,933 --> 00:05:39,099
because then we're not actually binning our data. It's just showing each and every single value

89
00:05:39,100 --> 00:05:45,300
that we have. And this doesn't really look like a shape that we can use.

90
00:05:45,300 --> 00:05:45,733


91
00:05:45,733 --> 00:05:50,399
Whereas if we set it to something more like between 30 and 50, suddenly we

92
00:05:50,400 --> 00:05:54,633


93
00:05:54,633 --> 00:06:00,566
can see the shape of our data much more clearly. So just be careful when you're setting your bins that you're

94
00:06:00,566 --> 00:06:00,966


95
00:06:00,966 --> 00:06:07,332
doing enough, but not too many. And if you're confused about whether or not you've got too many or enough reach out to your instructor

96
00:06:07,333 --> 00:06:13,633
or you know talk to another student, talk to a teacher and see if you can get some feedback. All right,

97
00:06:13,633 --> 00:06:15,033


98
00:06:15,033 --> 00:06:18,033
that's it. Everybody have a great day.