A tabular representation of data table is k/a table which have rows and column

Why are bar charts plotted.

Frequency table is created in backhand from which bar graph is created having horizontal and
vertical variables on x and y axis

If you want to visualize a data after grouping your data into frequencies than you plot a bar graph

One is categorical and other column is continues variable

Pie chart

It also uses frequency count like bar graph here we use percentage occurrence of each variable that’s
why pie chart is created

So in a pie based visual from the percentages got above we create charts

Line chart

When dots are created and connected after frequency table is created line chart is created

So when to use bar and when to use line- to see anything trends we use line chart(i.e sometime data
is high sometimes low ex say there is table having sales as 10 in jan 20 in feb 30 in mar then 15 in
April and so on so in this data sales first increases then decreases further may increase or decrease
so we use line chart)

Histogram

Are like bar charts we group data into bins to plot it

Ex

Candi age

P1 21

P2 22

P3 30

P4 45

I can convert ages in bin like 15-25,26-40,41-55…..

Once bins are created frequency table is created according to the bins and graph is drawn

Like in above example for bin 15-22 p1 and p2 will come as 2, p3 will count in 26-40 and p4 in 41-55
and so on

Measure of central tendency(tells our data middle point)

 Mean(middle term of data)

Measure of spread(it tells from center how is our data distributed)

We find range in this(by subtracting largest number from smallest number)

Range tell us how our data is distribured(say if from subtraction of both that is too big so we can say
distribution is big else low)

IQR-Inter quartile Range

We measure distance b/w quarters of our data

Quartiles are of 4 type

1)Q1-25th percentile-how many percent people are behind you

2) Q2-50th percentile-how many percent people are behind you

3) Q3-75th percentile-how many percent people are behind you

4) Q4-100th percentile-how many percent people are behind you

Say you have total data of 18 people so to find q1

Q1=25*18/100 q2=50*18/100…

I will always divide my data into 2 parts the middle one will be 50th percentile that is q2(is equivalent
to median)

The 1st divided series middle point will give q1 and likewise we will get q3

IQR=q1-q3

If IQR is low it means data is close to median and if it is high tells your data is widely spread from
center

Outliers

In small data when you see a number that is deviated from other numbers is k/a outlier

Say your data is 1,2,4,7,9,10,101

101 will be outlier

So say for data we find

mean=19.14

median=7

mean and median values are different so there is no central tendency

outliers effect your mean majorly.

Mean=23/6=3.8….

Median=4+7/2=5.5

How to remove outliers

We can create an upper and lower fence in our data the values above these is outlier

To create these fence we will use IQR

So

q1= 2

q3= 10

IQR=10-2=8

Upper fence=q3+1.5IQR

Lower fence=q1-1.5IQR

So in this example

Upper fence=10+1.5*8=22(101 will be considered as outlier and therefore we can remove 101 from
our data)

Lower fence=2-1.5*8=-10

Five number summary

Minimum

Q1

Q2

Q3

MAXIMUM

Before finding out 5 number summary you need to remove outlier from the data

5 number summary is used in creating box plot

Is you data tightly clusted or open around you mean is k/a variance

So variance=((n1-mean)+ (n2-mean)+……)2/n(total no of elements)

SD=Variance under root

Distance b/w your mean from your given element

Variance and sd is for sample of data

Density curve
Fo creating density curve we first create histograms and we join the upper most point of the graphs
by a line and remove the bars below it is k/a frequency polygon (lower the bin more inaccurate
polygon)

When you create more bins it will create like a straight line from these point this is known as density
curve

Types of density curves

DC tells us distribution around the curve

1)symmetrical/normal/guassian distribution

2)skwed distrubition

a)left skewed-tail is at left i.e majority of data is coming to left and less on other points

b)right skewed-tail is at right i.e majority of data is coming to left and less on other points

whenever you have skewed distribution your mean median mode will be very different

skewed distribution is mainly due to outliers in the data

in normal distribution to find center of tendency you use mean

whereas in skewed one you use median

Z-score

The area under normal distribution is always 100

Emperical rule is followed in Normal distribution

We divide our distribution in three parts 1st line 68% 2nd line at 95% and last line at 99.7%

These points come frome SD points(SD1 IS 68%,SD2 IS 95%,SD3 IS 99.7%)

Example data

1,2,3,4,5,6,7,8,9

Mean=5
SD=2.73
WE WILL CREATE A PLOT FROM THIS FIRST WE WILL CREATE A CENTER POINT AS 5 THEN FROM 5 WE
WILL ADD 2.73=7.73 AND SUBTRACT 2.73 I.E 2.27 WE GET FIRST 68% POINTS AFTER THAT WE WILL
ADD 5+2(2.73) AND 5-2(2.73) WE WILL GET 95% TWO POINTS AND LIKEWISE ADD 5+3(2.73) AND 5-
3(2.73) WE WILL GET 99.7% TWO POINTS
What is use of z score
It is used to understand that how far is the datapoint from mean measured in terms of standard
deviation
Z score=(data point-mean)/standard deviation
Z score tells us total percentage data below specified data
Example data
Mean=8
SD=1
Data point=9.5
Therefore your z score=1.5
We go to z table on google there we find out 1.50 zscore corresponds to .9332 that implies that
where my z score was coming as 1.5 all the values below tht point are 93.32% of all values held and
above that 100-93.32 is no of data above that 1.5 that is 6.68