Afrikaans
Akan
Albanian
Amharic
Armenian
Azerbaijani
Basque
Belarusian
Bemba
Bengali
Bihari
Bosnian
Breton
Bulgarian
Cambodian
Catalan
Cebuano
Cherokee
Chichewa
Chinese (Simplified)
Chinese (Traditional)
Corsican
Croatian
Czech
Danish
Dutch
English
Esperanto
Estonian
Ewe
Faroese
Filipino
Finnish
French
Frisian
Ga
Galician
Georgian
German
Greek
Guarani
Gujarati
Haitian Creole
Hausa
Hawaiian
Hebrew
Hindi
Hmong
Hungarian
Icelandic
Igbo
Indonesian
Interlingua
Irish
Italian
Japanese
Javanese
Kannada
Kazakh
Kinyarwanda
Kirundi
Kongo
Korean
Krio (Sierra Leone)
Kurdish
Kurdish (Soranรฎ)
Kyrgyz
Laothian
Latin
Latvian
Lingala
Lithuanian
Lozi
Luganda
Luo
Luxembourgish
Macedonian
Malagasy
Malay
Malayalam
Maltese
Maori
Marathi
Mauritian Creole
Moldavian
Mongolian
Myanmar (Burmese)
Montenegrin
Nepali
Nigerian Pidgin
Northern Sotho
Norwegian
Norwegian (Nynorsk)
Occitan
Oriya
Oromo
Pashto
Persian
Polish
Portuguese (Brazil)
Portuguese (Portugal)
Punjabi
Quechua
Romanian
Romansh
Runyakitara
Russian
Samoan
Scots Gaelic
Serbian
Serbo-Croatian
Sesotho
Setswana
Seychellois Creole
Shona
Sindhi
Sinhalese
Slovak
Slovenian
Somali
Spanish
Spanish (Latin American)
Sundanese
Swahili
Swedish
Tajik
Tamil
Tatar
Telugu
Thai
Tigrinya
Tonga
Tshiluba
Tumbuka
Turkish
Turkmen
Twi
Uighur
Ukrainian
Urdu
Uzbek
Vietnamese
Welsh
Wolof
Xhosa
Yiddish
Yoruba
Zulu
1
Today we're going to continue working our data mining workbook which contains the bank churn data set
2
.
3
However we will start looking at a bit of a different rule.
4
Subtle looking at our daughter from a bit of a different angle.
5
Let's create a new work tap and this one will call and well we won't call it anything yet.
6
What we want to visualize here is we want to understand how our customers are distributed by age so
7
our old customers predominately older or younger what age groups are they in.
8
And so on.
9
And as you can see we already have an age variable here.
10
So the first thing that comes to mind is let's say age and drag it into columns.
11
And now let's take number of records and drag it into rows.
12
And now we have one big dot over here at the top.
13
Why do we have one big dog.
14
Because well what happened is Tablo took some of Age of all of your customers and it got just over 380000
15
years and then it took some more records so the toll on them are at it's 10000.
16
Before we continue I'm just going to fix up the format here so we can see better.
17
So you can see here three eighty thousand not for 90000 and 10000 customers.
18
So how do we avoid doing that.
19
Well in my course on Tablo I talk about aggregation and granularity.
20
We won't go into details about that yet.
21
I'll just show you how to avoid this happening.
22
So if you go to some age and you just change this from a measure to a dimension then what will happen
23
is it's not aggregating age anymore it's using it as a dimension.
24
And as you can see you've got this nice shape of a curve.
25
So basically it just tells you how many customers at every given age they are and it's uploading it
26
with a line.
27
So that's good but the problem here is that this is a nice data set so I'm going to Right-Click and
28
I'll show you the data such if you Right-Click a data set up here and click your data and you go in
29
here you will see the details of your data.
30
So there we go.
31
There's age.
32
So here you can see that this is a nice data set in the sense that age has been around for you in a
33
way it's not good because you don't have the details how exactly how old the person is is he for 29
34
years between nine and a half or Z 29 and 11 months and a couple of days old.
35
So you don't have that decimal point you can tell how exactly all the Is and therefore you've lost some
36
information but on the other hand you can't do anything about it.
37
If this is how the data said was provided to well then you've got to work on it and you might as well
38
just enjoy it because that means it's a nice data set and you have one less thing to worry about you
39
don't have the decimal point to worry about.
40
So what does that say to us.
41
Well that says to us that yes in this case because they're rounded you Tablo will be able to group customers
42
for instance there's a person who's 36 and there is a person who's 36 and there's a person who is 36
43
and there's a couple more people who are 36 and then you go to this chart and then you find 46.
44
Let me just close this and you find 36 and you find that there's is 456 people that have thirty six
45
in that column.
46
What you will find is with other data sets when it's not rounded for you then you'll have people who
47
are three six and a half.
48
You'll have the people of thirty six point to thirty six point three and that will cause this charge
49
to go to create very strange patterns because you might have a lot of people who have three six and
50
a half but you might have very few people or three 6.6 so it will be very spiky and you don't want that
51
in your body you can really see a spike over here.
52
This is not normal right.
53
So 30 29 and then 31 so 348 people foreign for people on this island.
54
Just you know just so happened maybe these people are about to turn 30 or these people just turned 31
55
and therefore you have this spike and you'll see a lot more of them.
56
If your daughter is I would say more proper in the sense it's more precise.
57
So you have to find a way around that.
58
And we have to find a way to visualize our distribution of age regardless of the way our daughter is
59
presented to us.
60
And this is where we can use bins. This is where bin's come in and this is where we start thinking about
61
converting our numeric variables to categorical variables in two dimensions.
62
So let's start by having a look at that.
63
I'm going to Right-Click age and what I want to do is I want to group people into bands.
64
So I want to group people into for instance 5 year bans anybody between 20 and 25 will go into one group
65
of 25 to 30 will go to the next group and so on.
66
Well it's very easy to do it and Tablo all you have to go do is right click your variable and then here
67
select Create and then select bins.
68
And here you want an age been so we don't want size to be 10 one size to be five.
69
So every five years there'll be a new bin will click OK.
70
And here you can see Age state as a measure.
71
But now we have a dimension which is also age but it's called Age been.
72
And now what we'll do is we'll take age pin and will drag it in to replace age.
73
Well actually I need to take age out and then put age and been there because one's a dimension that
74
is a measure and what you can see right away is a nice distribution just like the ones we are used to
75
seeing in economics and in mathematics because what happened here.
76
We've grouped our records by category or by these bins and the bins are acting as categories so just
77
like let's say gender or the country that people live in a similar way so it's not a recognisable is
78
no longer recognising these as numbers these are categories so people in this category are between 20
79
and 25 people in this category between 25 and 30 by the way.
80
It includes the starting numbers so starting from 25 up to the last number before 30 in our case it's
81
29 so 25 29 30 to 34 35 to 39 and so on and right away you can see how many people they are in each
82
of these categories.
83
So what else would we want to do here.
84
Maybe let's give it a color so it looks a bit more pleasant.
85
Let's take some number records hold down control and ragged on to color.
86
Personally I use green for money.
87
So let's use a different color.
88
How about we go with kind of a brownish color.
89
Something that apply and maybe we want to add a label as well so we don't have to visually check every
90
single time where this is sitting.
91
So we'll go and use some number of records as a label as well.
92
And now you can see that.
93
Exactly how many people you have in each of these categories.
94
So and this approach allows you to assess your daughter a bit more so now you understand from this sample
95
which is quite a significant sample 10000 people in the sample So it's very representative of the overall
96
population of this bank and what you can see is that most people are age 35 to 40.
97
Quite a few people between 30 to 40 so basically this is telling you what is the demographic of the
98
bank like they are.
99
Let's say more younger people say in their 30s that most customers are presented by age groups between
100
30 and 45.
101
So you have the most of your customers are in those age groups.
102
And in fact if you want to see the percentages you can easily convert this so you can just go here and
103
add table calculation like we did previously.
104
You can even at a quick table calculation in this case.
105
So you just need a percent of total.
106
And there you go so you have the exact percentages.
107
This gives you 43 plus 43 plus 16 gives you 59 or 59 plus.
108
There's like a percentage so that gives you 60 percent 60 percent of your customers are in these H-back
109
age bands and the rest there's like a long tail.
110
So this is a right skewed distribution.
111
So that is a thing to remember as well so maybe another tip that when that skewed means where the tail
112
is going if the tells going to the right then it's a right skewed.
113
This is not a left skewed a lot of people get it wrong left skewed would be if the tail was going to
114
the left so that's a statistical term to keep in mind it's a rice food distribution got a long tail
115
to the right you and have customers at 90 years old.
116
And so what are we going to do now.
117
That's that's pretty much what we want to do with this chart.
118
I'll just say one quick thing you can change the bins very easily.
119
You can just go to age and then you go to Edit.
120
And here let's say we want a 10 year band.
121
See right away it's changed and now you can see that you got most customers between 30 and 40 and then
122
40 to 50 you have 26 and so rich controls at that.
123
And that's all we wanted to do here.
124
The last thing you might want to do is change this access because it's not consistent with your percentage
125
you might want to change a percentage but that's details we're just investigating so we won't be doing
126
it right now.
127
And that's all for today.
128
I look forward to seeing you next time.
129
And until then happy analyzing
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.