DESCRIPTIVE STATISTICS USING R

UNIT 4
IMPORT → VISUALIZE → DESCRIBE
- IMPORT CSV/EXCEL FILES.
- USE R TO CREATE BASIC CHARTS.
- UNDERSTAND DESCRIPTIVE STATISTICS.
INSTALLING PACKAGES AND LOADING THEM

 # Install if not already installed

 [Link]("tidyverse") # For data manipulation and visualization
 [Link]("readxl") # For Excel file import
 [Link]("dplyr") # For advanced data manipulation

 # Load libraries
 library(tidyverse)
 library(readxl)
 library(dplyr)
IMPORTING FROM EXCEL AND EXPLORING THE FILE
 library(readxl)
 CreditLimit <- read_excel(“path of the file") #make sure to use forward slashes “/”
 View(CreditLimit)
 #check the first 6 rows
 head(CreditLimit) # by default, the first 6 rows are checked
 #check the last 6 rows
 tail(CreditLimit)
 #check the summary
 summary(CreditLimit)
IMPORTING FROM EXCEL AND EXPLORING THE FILE
 #check the summary of a variable
 summary(CreditLimit$Income)
 #check dimensions
 dim(CreditLimit)
 #check the number of variables
 length(CreditLimit)
 #check the list of variables
 attributes(CreditLimit)
 #check the data type of variables
 class(CreditLimit)
IMPORTING FROM EXCEL AND EXPLORING THE FILE
 #check the internal storage of the class
 typeof(CreditLimit)
 #check number of rows
 nrow(CreditLimit)
 #check number of columns
 ncol(CreditLimit)
 # check column names
 colnames(CreditLimit)
 glimpse(data) # Tidyverse alternative to str()
 #check for missing values
 sum([Link](CreditLimit)) #counts total missing values
 colSums([Link](CreditLimit)) # missing values per column
DATA VISUALIZATION WITH R
HISTOGRAM

• A histogram is a graphical representation of the distribution of numerical data.

• It uses continuous bars (bins) to show the frequency of data within specific ranges.
• Ideal for visualizing the distribution of a continuous variable (e.g., income).
• Helps identify skewness, peaks, and gaps in the data.
HISTOGRAM

 # Base R Histogram of Income distribution

 hist(CreditLimit$Income,
 main = "Income Distribution",
 xlab = "Income ('000)",
 col = "lightblue",
 border = "black",
 breaks = 10)
 Histogram shows the frequency distribution of Income
 breaks defines the number of bins
BAR PLOT

• A bar plot (or bar chart) uses rectangular bars to represent data values.
• The height or length of each bar is proportional to the magnitude of the variable it
represents.
• Best suited for categorical data (e.g., regions, marital status).
• Helps compare counts or averages across different groups.
BAR PLOT
 # suppose we want to show the number of customers by Marital Status
 # let’s first create a frequency table counting the occurrences of each category in the Married column
 marital_count <- table(CreditLimit$Married)
 # the result is a named vector with counts for Yes and No
 # now let’s construct a bar plot
 barplot(marital_count,
 main = "Customer Count by Marital Status",
 xlab = "Marital Status",
 ylab = "Count",
 col = "skyblue",
 border = "black")
BAR PLOT – ANOTHER EXAMPLE
 # Base R Bar plot showing average Credit Limit by Region
 avg_limit <- tapply(CreditLimit$’credit Limit’, CreditLimit$Region, mean)
 # tapply() calculates the mean by region.
 barplot(avg_limit,
 main = "Average Credit Limit by Region",
 xlab = "Region",
 ylab = "Average Credit Limit",
 col = "skyblue",
 border = "black")
 Bar plots display aggregated values by category.
BOX PLOT OR BOX-AND-WHISKER PLOT

• A box plot displays the five-number summary of a dataset:

• Minimum,
• 1st quartile (Q1),
• Median (Q2),
• 3rd quartile (Q3),
• Maximum.

• It also highlights outliers and the spread of the data.

• Useful for comparing distributions between groups.
• Helps detect outliers and skewness.
BOX PLOT

 # Simple box plot for Income distribution

 boxplot(CreditLimit$Income,
 main = "Box Plot of Income",
 ylab = "Income",
 col = "lightblue",
 border = "darkblue")
BOX PLOT – ANOTHER EXAMPLE

 # suppose we want to construct a Box plot comparing credit limit by Marital Status
 boxplot(CreditLimit$’credit Limit’ ~ CreditLimit$Married, # ~ groups by Married
status
 main = "Credit Limit by Marital Status",
 xlab = "Marital Status",
 ylab = "Credit Limit",
 col = c("pink", "lightgreen"))
LINE PLOT WITH SINGLE SERIES OF DATA
 # create a vector l1<- c(7,12,28,3,41)
 l1
 # plot l1 using points plot(l1, type = “p”)
 # plot l1 using lines plot(l1, type = “l”)
 # plot l1 using both points and lines plot(l1, type = “o”)
 # plot a vector using both points and lines, with colour
 l2<- c(1,2,8,13,40)
 plot(l2, type = “o”, col=“red”)
 # plot a vector using both points and lines with colour, heading, label of the x & y axis
 l3<- c(5,2,11,7,20,15,22,17,25)
 plot(l3, type = “o”, col=“green”, main=“Line Plot”, xlab=“points”, ylab=“Frequency”)
LINE PLOT WITH MULTIPLE SERIES OF DATA
 # let variable 't' represent time
 t<-0:10
 #variable 'z' showing quantity that is decreasing in time
 z<-exp(-t/2)
 #variable 'w' that is increasing with time
 w<-0.1*exp(t/3)
 #plot t and z
 plot(t, z, type =“l”)
 #plot t and z with colour red, line width 3, label of x & y axes time and concentration
 plot(t, z, type ="l", col="red", lwd="3", xlab="Time", ylab="Concentration")
LINE PLOT WITH MULTIPLE SERIES OF DATA
 #plot t and w
 plot(t, w, type =“o”)
 #plot t and w with colour green, line width 4, label of x & y axes time and
concentration
 plot(t, w, type =“o”, col=“green”, lwd=“4”, xlab=“Time”, ylab=“Concentration”)
 # we can add a line to an existing plot using lines() function (superimposes)
 lines(t, z, col=“red”, lwd=“3”)
 #add title 'Exponential Growth and decay’
 title("Exponential Growth and decay")
LINE CHART

• A line graph uses points connected by lines to represent data trends

over time or sequential values.
• The x-axis typically represents the time or order variable, while the y-
axis shows the values.
• Ideal for visualizing trends or changes over time.
• Shows continuous data progression (e.g., customer ID vs. credit limit).
LINE CHART

 # Base R Line plot of Bank Balance vs Age

 plot(CreditLimit$Age, CreditLimit$`Bank Balance (‘000)`,
 col = “blue”,
 main = “Bank Balance vs Age”,
 xlab = “Age”,
 ylab = “Bank Balance ('000)”
 type=“o”)
SCATTER PLOT

• A scatter plot displays individual data points on a two-dimensional plane.

• It shows the relationship between two continuous variables, where:
• X-axis → Independent variable.
• Y-axis → Dependent variable.

• Used to identify correlations or patterns between two variables.

• Helps detect linear or non-linear relationships.
SCATTER PLOT
 # Base R Scatter plot of Income vs Credit Limit
 plot(CreditLimit$Income, CreditLimit$`credit Limit`,
 main = "Income vs Credit Limit",
 xlab = "Income ('000)",
 ylab = "Credit Limit",
 col = "purple",
 pch = 19) # pch defines point shape
 # Create a plot displaying all pch types
 plot(1:25, rep(1, 25), pch = 1:25, cex = 2) # Increase point size
pch Value Symbol Description
0 □ Square outline
1 ○ Circle outline
2 △ Triangle outline (up)
3 + Plus
4 × Cross
5 ◇ Diamond outline
6 △ (filled) Triangle (up, filled)
7 ▽ Triangle outline (down)
8 Asterisk
9 ◇ (filled) Diamond (filled)
10 ⬥ Solid circle with cross
11 ⬥ Star-like shape
12-25 ⬤ Varied filled and unfilled shapes
DESCRIPTIVE STATISTICS
SIMPLE EXAMPLE FIRST..

 #create a vector
 x<-c(3,7,5,13,20,23,39,23,40,23,14,12,56,23)
 #calculate mean of x
 mean(x)
 #calculate median of x
 median(x)
 #calculate mode of x
 #create a frequency table and store in a variable t1
 t1<-table(x)
 t1
 max(t1) # finds maximum value
 t1==max(t1) # finds position of the maximum value
 names(t1[t1==max(t1)]) # extracts the names corresponding to the maximum
value
 mode_x <- [Link](names(t1[t1==max(t1)]))
 mode_x
FURTHER PRACTICE..

 #create a vector
 y<-c(3,7,5,13,20,20,39,23,40,23,14,12,56,23,20)
 #calculate mean, median and mode of y
 #create a vector
 z<-c(3,7,5,13,20,20,39,23,25,25,40,23,14,12,56,23,20,25)
 #calculate mean, median and mode of z
 # similarly find mean and median for credit Limit column in CreditLimit data
MULTIMODAL DATA
 In case of multimodal data, better way is as below (and this is standardized for all
datasets)
 find_mode <- function(x) {
freq_table <- table(x)
mode <- [Link](names(freq_table[freq_table == max(freq_table)]))
return(mode)
}
 find_mode(CreditLimit$`credit Limit`)
 find_mode(x)
 # Find mode for data frame
 # first create a data frame
 df <- [Link](
x = c(1, 4, 4, 5, 6, 7, 10, 12),
y = c(2, 2, 3, 3, 4, 5, 11, 11),
z = c(8, 9, 9, 9, 10, 13, 15, 17)
)
 print(df)
 # now apply the function to calculate mode for each column of the data frame above (hint: apply family
of functions)
MEASURES OF DISPERSION
 # Range, variance, and standard deviation of Income
 [Link]<-range(CreditLimit$Income) # Min and Max
 [Link]
 [Link]<-[Link][2] – [Link][1]
 Or
 [Link] <- range(CreditLimit$Income)[2] - range(CreditLimit$Income)[1]
 Or max(CreditLimit$Income) – min(CreditLimit$Income)
 Or we could create a function for the same
 var(CreditLimit$Income) # Variance
 sd(CreditLimit$Income) # Standard deviation
 IQR(CreditLimit$Income) # Inter quartile range
COVARIANCE AND CORRELATION
 # simple example first
 u<-c(1,2,3,4,5,6,7,8,9,10)
 v<-c(10,9,8,7,6,5,4,3,2,1)
 w<-c(2,4,6,8,10,12,14,16,18,20)
 # find covariance and correlation between u and v
 cov(u,v) cor(u,v)
 # find covariance and correlation between u and w
 cov(u,w) cor(u,w)
 # calculate covariance and correlation between Income and Credit Limit
 cov(CreditLimit$Income, CreditLimit$`credit Limit`) # Covariance
 cor(CreditLimit$Income, CreditLimit$`credit Limit`) # Correlation
 Calculate covariance using Pearson method
 Calculate covariance using Spearman method
 Calculate covariance using Kendall method
COEFFICIENT OF DETERMINATION
 The Coefficient of Determination (R²) is the Square of the Coefficient of Correlation (r).
 For simple linear regression (with one independent variable), the relationship between the two
is:
 R2 = (r)2
 r is the Pearson correlation coefficient, which measures the strength and direction of the
linear relationship between two variables.
 R2 is the coefficient of determination, which indicates the proportion of variance in the
dependent variable explained by the independent variable.
 # example
 x<-c(1,3,5,10)
 y<-c(2,4,6,20)
 r_squared<-cor(x,y)^2
 r_squared
COEFFICIENT OF DETERMINATION
 # calculating R2 using linear regression model
 # same sample data
 x <- c(1, 3, 5, 10)
 y <- c(2, 4, 6, 20)
 # Create a linear model
 model <- lm(y ~ x)
 # Extract R-squared
 r_squared <- summary(model)$[Link]
 # Print R-squared
 print(r_squared)
 cat("Coefficient of Determination (R²):", round(r_squared, 4))
COEFFICIENT OF DETERMINATION (R²)
• R² measures how well the model explains the variability of the data.
• Values range from 0 to 1 (higher values indicate a better fit).
 # applying the linear regression model method to credit Limit and Income
 model <- lm(CreditLimit$`credit Limit` ~ CreditLimit$Income)
 # R-squared value
 r_squared<-summary(model)$[Link]
 cat("Coefficient of Determination (R²):", round(r_squared, 4))
 Explanation:
 R² ≈ 0.6274 means that 62.74% of the variance in the Credit Limit is explained by Income.
 The remaining 37.26% is due to other factors not included in the model.
PRACTICING USING ANOTHER FILE…
IRIS DATASET IN R
EXPORTING FROM R

 Let’s first export the dataset to a .csv file and then import it
 # write the dataset into a data object
 data<-iris
 [Link](data, "C:/Users/payal/Documents/[Link]", [Link]=FALSE)
 # check if the file has been saved at the specified location
 [Link]("C:/Users/payal/Documents/[Link]“)
IMPORTING INTO R
 Write an R script to import the [Link] file into a data frame.
 Iris <- [Link](“path to the file”)
 How do you check if the file was imported correctly?
 Check the first few rows.
 What function can you use to view the first 10 rows of the dataset?
 head(Iris,10)
 How do you check the structure of the dataset in R?
 str(Iris) OR glimpse(Iris)
 How do you verify if there are missing values in the dataset?
 sum([Link](Iris))
EXPLORING THE DATASET
 How do you find the number of rows and columns in the dataset?
 dim(Iris)
 How do you extract only the first and last 5 rows of the dataset?
 head(Iris, 5)
 tail(Iris,5)
 Write R code to display the column names of the dataset.
 colnames(Iris)
 How do you determine the data type of each column?
 sapply(Iris, class)
 Write an R command to get a summary of the dataset.
 summary(Iris)
VISUALIZATIONS USING BASE R
 Create a histogram for the [Link] column, adding colour, labels, and title.
 hist(Iris$[Link]……..)
 Generate a boxplot for [Link], adding all other elements.
 boxplot(Iris$[Link], ……)
 Create a scatter plot between [Link] and [Link], fully detailed.
 plot(Iris$[Link], Iris$[Link], ….)
 Differentiate species using different colors in the scatter plot created above.
 plot(Iris$[Link], Iris$[Link], col = Iris$Species, pch = 19)
 Create a bar plot showing the count of each species in the dataset.
 barplot(table(Iris$Species),….)
MEASURES OF CENTRAL TENDENCY
 Write R code to calculate the mean, median, and mode of [Link].
 mean(Iris$[Link])
 median(Iris$[Link])
 find_mode(Iris$[Link]) # user-defined function
 How do you handle missing values when computing the mean?
 mean(Iris$[Link], [Link]=TRUE)
 Calculate the median for the [Link] column.
 median(Iris$[Link])
 Write a function to compute the mode of any given numeric column.
 Write the find_mode function.
 How do you find the most frequently occurring value in [Link]?
 find_mode(Iris$[Link])
MEASURES OF DISPERSION

 Compute the range of the [Link] column.

 Use the range function
 Write R code to calculate the variance of [Link].
 var(Iris$[Link])
 How do you find the standard deviation of [Link]?
 sd(Iris$[Link])
 What does a higher standard deviation indicate in terms of data spread?
 Create a boxplot to visualize the spread of [Link].
 boxplot(Iris$[Link]……)
COVARIANCE AND CORRELATION
 Compute the covariance between [Link] and [Link].
 cov(Iris$[Link], Iris$[Link])
 Compute the correlation coefficient between [Link] and [Link].
 cor(Iris$[Link], Iris$[Link])
 How do you interpret a correlation coefficient close to 1, 0, and -1?
 Perfect positive, no correlation, perfect negative
 Write an R function that takes two numerical columns and returns their correlation.
 correlation_fn <- function(x, y) {
 cor(x, y)
 }
 correlation_fn(iris$[Link], iris$[Link])
COEFFICIENT OF DETERMINATION (R²)
 Write R code to perform a linear regression between [Link] (dependent variable) and
[Link] (independent variable).
 model<-lm(Iris$[Link] ~ Iris$[Link])
 Extract the R-squared value from the linear regression model.
 r_squared<-summary(model)$[Link]
 Interpret an R² value of 0.85 in terms of model fit.
 How do you check the summary statistics of the linear regression model in R?
 summary(model)
 Create a scatter plot with a regression line for [Link] vs [Link].
 plot(data$[Link], data$[Link], main = "Petal Width vs Petal Length", xlab = "Petal
Width", ylab = "Petal Length", col = "blue", pch = 19)
 abline(model, col = "red", lwd = 2) # to add reference lines to an existing plot
ASSIGNMENT

 Import ‘Alpha Advertising Ltd.’ dataset into R.

 Draw a histogram of the total sales. Give a suitable title to the histogram, label both its axes, and colour it also.
 Draw a box plot of sales from newspaper. Give a suitable title to the plot, label the axes, and color the box plot,
with border.
 Draw a line plot with sales from radio and sales from T.V. add a legend also.
 Calculate the measures of centrality of each type of advertising mode.
 Calculate the measures of dispersion of each type of advertising mode.
 Calculate correlation between sales from TV and sales from newspaper.