![Measures of Central Tendency
mean(salary$ba,na.rm=T)
[1] 17209.74
mean(salary$ba )
[1] NA
# Mean
Using na.rm=T excludes the missing values from the mean
mean() gives mean of the variable.
median(salary$ba,na.rm=T)
[1] 16230
# Median
median(), gives median of the variable.
trimmed_mean<-mean(salary$ba,0.10,na.rm=T)
trimmed_mean
[1] 16879.09
Trimmed mean is a useful measure in the presence of outliers.
Using 0.10 in the mean(), excludes 10% observations from each side of
the data from the mean
names(sort(-table(salary$ba)))[1]
[1] "10940"
# Mode
There is no direct function in R for Mode. Hence, we
first obtain counts for each value using table() and
sort() to find the highest frequency.
name() displays the value of the first (highest
frequency) observation. 5](https://image.slidesharecdn.com/basicanalysisusingr-170925105146/85/Basic-Analysis-using-R-5-320.jpg)
![6
r<-range(salary$ba,na.rm=T)
r
[1] 10940 29080
# Range / Inter Quartile Range
range() gives minimum and maximum values of
that variable
diff(r)
[1] 18140
diff() calculates difference between all values of that
vector
IQR(salary$ba,na.rm=T)
[1] 9070 IQR() gives the Inter-Quartile range of the variable
Measures of Variation](https://image.slidesharecdn.com/basicanalysisusingr-170925105146/85/Basic-Analysis-using-R-6-320.jpg)
![var(salary$ba,na.rm=T)
[1] 17301567
7
sd(salary$ba,na.rm=T)
[1] 4159.515
# Standard Deviation/ Variance
sd() gives standard deviation of the variable
var() gives variance of the variable
cv<-sd(salary$ba,na.rm=T)/mean(salary$ba,na.rm=T)
cv
[1] 0.2416954
# Co-efficient of Variation
There is no standard function for CV in R. Hence we
calculate it by definition
Measures of Variation](https://image.slidesharecdn.com/basicanalysisusingr-170925105146/85/Basic-Analysis-using-R-7-320.jpg)
![Skewness and Kurtosis
kurtosis(salary$ba,na.rm=T,type=2)
[1] 0.4996513
8
skewness(salary$ba,na.rm=T,type=2)
[1] 0.9033507
# Skewness
skewness() gives skewness of the variable.
type=2 uses moment based formula
library(e1071) Using package “e1071” in R is the easiest way
to find skewness and kurtosis
# Kurtosis
kurtosis() gives kurtosis of the variable.
type=2 uses moment based formula
f<-function(x)
c(skew=skewness(x,na.rm=T,type=2),kurt=kurtosis(x,na.rm=T,type=2))
aggregate(ba~Grade,data=salary,FUN=f)
Grade ba.skew ba.kurt
1 GR1 0.85500651 0.62939005
2 GR2 0.08682743 0.34852383
# Skewness and Kurtosis by Grade](https://image.slidesharecdn.com/basicanalysisusingr-170925105146/85/Basic-Analysis-using-R-8-320.jpg)
![Pearson Correlation Coefficient
cor(job$aptitude,job$job_prof)
# correlation
[1] 0.5144107
cor() gives the Pearson Correlation Coefficient of the
two variables mentioned
15](https://image.slidesharecdn.com/basicanalysisusingr-170925105146/85/Basic-Analysis-using-R-15-320.jpg)
Basic Analysis using R
- 1. Basic Analysis Using R
- 3. Data Snapshot The data has 41 rows and 7 columns First_Name First Name Last_Name Last Name Grade Grade Location Location Function Department ba Basic Allowance ms Management Supplements Data Description basic_salary_R3 3
- 4. Describing Variable summary(salary) 4 salary<-read.csv(file.choose(), header=TRUE) #Importing Data #Checking the variable features using summary function First_Name Last_Name Grade Location Function Kavita : 2 Joshi : 2 GR1 :23 DELHI :17 FINANCE :13 Mahesh : 2 Shah : 2 GR2 :17 MUMBAI:21 SALES :15 Nishi : 2 Singh : 2 NA's: 1 NA's : 3 TECHNICAL:11 Priya : 2 Arora : 1 NA's : 2 Ajit : 1 Bhide : 1 Ameet : 1 Bhutala: 1 (Other):31 (Other):32 ba ms Min. :10940 Min. : 2700 1st Qu.:13785 1st Qu.:10450 Median :16230 Median :12420 Mean :17210 Mean :11939 3rd Qu.:19305 3rd Qu.:14200 Max. :29080 Max. :16970 NA's :2 NA's :4 summary() gives descriptive measures for numeric variable
- 5. Measures of Central Tendency mean(salary$ba,na.rm=T) [1] 17209.74 mean(salary$ba ) [1] NA # Mean Using na.rm=T excludes the missing values from the mean mean() gives mean of the variable. median(salary$ba,na.rm=T) [1] 16230 # Median median(), gives median of the variable. trimmed_mean<-mean(salary$ba,0.10,na.rm=T) trimmed_mean [1] 16879.09 Trimmed mean is a useful measure in the presence of outliers. Using 0.10 in the mean(), excludes 10% observations from each side of the data from the mean names(sort(-table(salary$ba)))[1] [1] "10940" # Mode There is no direct function in R for Mode. Hence, we first obtain counts for each value using table() and sort() to find the highest frequency. name() displays the value of the first (highest frequency) observation. 5
- 6. 6 r<-range(salary$ba,na.rm=T) r [1] 10940 29080 # Range / Inter Quartile Range range() gives minimum and maximum values of that variable diff(r) [1] 18140 diff() calculates difference between all values of that vector IQR(salary$ba,na.rm=T) [1] 9070 IQR() gives the Inter-Quartile range of the variable Measures of Variation
- 7. var(salary$ba,na.rm=T) [1] 17301567 7 sd(salary$ba,na.rm=T) [1] 4159.515 # Standard Deviation/ Variance sd() gives standard deviation of the variable var() gives variance of the variable cv<-sd(salary$ba,na.rm=T)/mean(salary$ba,na.rm=T) cv [1] 0.2416954 # Co-efficient of Variation There is no standard function for CV in R. Hence we calculate it by definition Measures of Variation
- 8. Skewness and Kurtosis kurtosis(salary$ba,na.rm=T,type=2) [1] 0.4996513 8 skewness(salary$ba,na.rm=T,type=2) [1] 0.9033507 # Skewness skewness() gives skewness of the variable. type=2 uses moment based formula library(e1071) Using package “e1071” in R is the easiest way to find skewness and kurtosis # Kurtosis kurtosis() gives kurtosis of the variable. type=2 uses moment based formula f<-function(x) c(skew=skewness(x,na.rm=T,type=2),kurt=kurtosis(x,na.rm=T,type=2)) aggregate(ba~Grade,data=salary,FUN=f) Grade ba.skew ba.kurt 1 GR1 0.85500651 0.62939005 2 GR2 0.08682743 0.34852383 # Skewness and Kurtosis by Grade
- 9. Generating Frequency Tables 9 freq<-table(salary$Location, salary$Grade) freq # Frequency Tables GR1 GR2 DELHI 11 6 MUMBAI 10 10 table() gives the frequency of counts of the two variables mentioned. prop.table(freq) # Percentage Frequency Tables GR1 GR2 DELHI 0.2972973 0.1621622 MUMBAI 0.2702703 0.2702703 prop.table() gives the frequency expressed as percentage of total count. (,1) expresses the frequency as percentage of row count whereas (,2) would express it as percentage of column count
- 10. Generating Frequency Tables 10 table1 <- table(salary$Location, salary$Grade, salary$Function) ftable(table1) # Three Way Frequency Table FINANCE SALES TECHNICAL DELHI GR1 4 4 3 GR2 2 1 3 MUMBAI GR1 2 4 2 GR2 3 4 3 ftable() gives the frequency of counts of the three variables mentioned in one table itself
- 11. Generating Cross Tables 11 install.packages("gmodels") library(gmodels) # Installing package – “gmodels” # Frequency Table using ‘gmodels’ package CrossTable(salary$Grade,salary$Location, prop.r=FALSE, prop.c=FALSE) CrossTable() gives the frequency of counts of the two variables mentioned prop.r= removes the row proportion from the output and prop.c= removes the column proportion
- 13. Data Snapshot The data has 25 rows and 6 columns empno Employee Number aptitude Aptitude Score of the Employee testofen Test of English tech_ Technical Score g_k_ General Knowledge Score job_prof Job Proficiency Score Data Description job_proficiency_R3 13
- 14. Scatter Plot 14 plot(job$aptitude,job$job_prof,col="red") # scatterplot job<-read.csv(file.choose(),header=T) plot() gives a scatterplot of the two variables mentioned. col= provides color to the points • Study the correlation between Aptitude and Job Proficiency.
- 15. Pearson Correlation Coefficient cor(job$aptitude,job$job_prof) # correlation [1] 0.5144107 cor() gives the Pearson Correlation Coefficient of the two variables mentioned 15
- 16. attach(job) plot(aptitude,job_prof, main="ScatterPlot with Regression Line", xlab="Aptitude ", ylab="Job Proficiency", pch=19) abline(lm(job_prof~aptitude), col="darkorange") ScatterPlot with Regression Line #Scatterplot with Regression Line 16 plot() in base R yields a different types of plots aptitude is one of the variable for plot job_prof is another variable to be plotted main= provides the user defined name of the chart. It has to be put in double quotes xlab= provides a user defined label for the variable on X axis ylab= provides a user defined label for the variable on Y axis pch= gives various shapes for the data points on the plot abline() in base R yields a different types of lines on plot lm() provides the liner regression line of the first variable mentioned on the second col= provides the color of the line plotted attach() is used to call the data in R with help of which in further . Codes specifying the data repetitively can be avoided.
- 18. Scatter Plot Matrix 18 pairs(~job_prof+aptitude+testofen+tech_+g_k_,data=job, main="ScatterPlot Matrix",col="darkorange") #ScatterPlot Matrix Output: pairs() in base R is used to plot pairwise comparison ~ each variable name to be plotted followed by a “+” sign needs to mentioned main= provides the user defined name of the chart.
- 19. SECTION 3 Data Visualisation Graphs in R 19
- 20. Data Snapshot The data has 1000 rows and 10 columns CustID Customer ID Age Age Gender Gender PinCode PinCode Active Whether the customer was active in past 24 weeks or not Calls Number of Calls made Minutes Number of minutes spoken Amt Amount charged AvgTime Mean Time per call Age_Group Age Group of the Customer Data Descriptiontelecom_R3 20
- 21. telecom1<-aggregate(Calls~Age_Group,data = telecom, FUN=sum) telecom1 telecom<-read.csv(file.choose(), header=TRUE) Simple Bar Chart #Importing Data #Creating a Matrix 21 Age_Group Calls 1 >45 128870 2 18-30 943187 3 30-45 798721 For plotting a bar chart in R, it is important that the variable to be plotted is in a vector or a matrix form. table() converts the variable in vector/matrix form
- 22. Simple Bar Chart 22 barplot(telecom1$Calls, main= "SIMPLE BAR CHART (Total Calls - Age Group)", names.arg = telecom1$Age_Group, xlab = "Age Group",ylab="Total Calls",col= "darkorange") #Simple Bar Chart – Total Calls barplot() in base R yields different types of bar chart main= provides the user defined name of the chart. It has to be put in double quotes names.arg= takes vector of names to be plotted below each bar or group of bars. xlab= provides a user defined label for the variable on X axis ylab= provides a user defined label for the variable on Y axis col= can be used to input your choice of color to the bars
- 24. Stacked Bar Chart 24 #Stacked Bar Chart barplot() in base R yields different types of bar chart legend=rownames() displays the legend on the graph output telecom2<-table(telecom$Gender,telecom$Age_Group) telecom2 >45 18-30 30-45 F 32 256 221 M 39 245 207 table() inputting two variables gives a matrix having their counts in each category barplot(telecom2,main="STACKED BAR CHART", xlab ="Age Group",ylab="No.of Customers", col=c("cadetblue","orange"),legend=rownames(telecom2))
- 26. Percentage Bar Chart 26 #Percentage Bar Chart barplot() in base R yields different types of bar chart telecom3*100 a vector or matrix for which the bar chart needs to be plotted. *100 would display percentage scale on x-axis telecom3<-prop.table(telecom2,2) telecom3 barplot(telecom3*100, main= "PERCENTAGE BAR CHART", xlab = "Age Group",ylab="No.of Customers", col = c("cadetblue","orange"),legend=rownames(telecom3)) >45 18-30 30-45 F 0.4507042 0.5109780 0.5163551 M 0.5492958 0.4890220 0.4836449 prop.table() helps us create data frame with percentage values (,2) gives percentage as per column count
- 28. Multiple Bar Chart 28 # Multiple Bar Chart barplot() in base R yields different types of bar chart beside=TRUE enables us to show the different class of the same bar one beside the other telecom4<-xtabs(Calls~Gender+Age_Group,telecom) telecom4 barplot(telecom4,main="MULTIPLE BAR CHART (Total Calls - Gender & Age Group)", xlab ="Age Group",ylab="No.of Calls", col=c("cadetblue","orange"),legend=rownames(telecom4), beside = TRUE) Age_Group Gender >45 18-30 30-45 F 58310 480235 408184 M 70560 462952 390537 xtabs() is used to cross tabulate the categories of more than one variables using another numeric variable which results in totals in each category
- 30. Pie Chart 30 telecom5<-aggregate(Calls~Age_Group,data = telecom, FUN=sum) telecom5$pct <- round(telecom5$Calls/sum(telecom5$Calls)*100) telecom5 #Pie Chart Here, we calculate the proportions for each category using formula pie() in base R yields a pie chart labels= provides a user defined label for the variable on X axis paste() labels each category using string values separated by commas col= can be used to input your choice of color to the bars pie(telecom5$Calls,labels = paste(telecom5$Age_Group,"(",telecom5$pct,"%)"), col=c("darkcyan","orange","yellowgreen"), main="Pie Chart with Percentage") Age_Group Calls pct 1 >45 128870 7 2 18-30 943187 50 3 30-45 798721 43
- 32. Box Plot 32 boxplot(telecom$Calls,data=telecom, main="BOX PLOT (Total Calls)", ylab="Total Calls",col="cadetblue3") #BoxPlot – Total Calls boxplot() in base R yields a different types of box chart data= calls the data out of which the variable needs to be plotted main= provides the user defined name of the chart. It has to be put in double quotes ylab= provides a user defined label for the variable on Y axis col= can be used to input your choice of color to the bars
- 34. Histogram 34 hist(telecom$AvgTime, breaks=12, main = "HISTOGRAM - Average Call Time", xlab = "Average Call Time", ylab = "No. of Customers", col="darkorange") #Histogram – Average Call Time hist() in base R yields a histogram breaks= specifes the width of each bar main= provides the user defined name of the chart. It is to be put in double quotes xlab= provides a user defined label for the variable on X axis ylab= provides a user defined label for the variable on Y axis col= can be used to input your choice of color to the bars
- 36. Density Plot 36 telecom6<-density(telecom$Amt) plot(telecom6, main="DENSITY PLOT - Amount",xlab="Amount") polygon(telecom6, col="yellowgreen") #Density Plot - Amount density() returns the density values of the variable plot() plots the line graph of the specified variable main= provides the user defined name of the chart. It is to be put in double quotes xlab= provides a user defined label for the variable on X axis polygon() shows the area covered under the curve col= can be used to input your choice of color in the polygon
- 38. Dot Plot 38 dotchart(telecom$Calls,cex=.7,main="DOT PLOT", xlab="Total Calls",col="red") #DotPlot dotchart() in base R yields a dotchart cex= specifies the amount by which plotting text and symbols should be scaled relative to default of 1 main= provides the user defined name of the chart. It has to be put in double quotes xlab= provides a user defined label for the variable on X axis col= can be used to input your choice of color in the dotchart
- 40. stem(telecom$Calls) Stem and Leaf Plot 40 #Stem and Leaf Plot in R Output: stem() in base R yields a stem and leaf chart
- 41. Heat Map 41 pheatmap(heatmap, display_numbers=TRUE,main="HEAT MAP", fontsize=12, fontsize_row=15, number_color="black") # Heat Map install.packages("pheatmap") library(pheatmap) #Installing and loading the package heatmap<-xtabs(Calls~Age_Group+Gender,data=telecom) #Creating a Matrix display_numbers=T shows the frequency of each cell main= provides the user defined name of the chart fontsize= inputs the font size of the main title fontsize.row= inputs the font size of the row labels number_color= inputs the color of the frequency shown in each cell pheatmap is the best package we can use to plot an effective Heat Map in R
- 43. THANK YOU! 43