PA Course 1 Exercise 2 with Answer Key

---

EXERCISE 2

BASIC STATISTICS AND DATA VISUALIZATION IN R

I. OBJECTIVES

At the end of this exercise, the participant must be able to:

1. generate and interpret descriptive statistics for variables and

2. generate and interpret graphs appropriate to the type of data in R.

II. METHODS

A. Using the data set patients.csv:

1. Load the data into the object ds1.

2. Generate appropriate descriptive summary statistics for all the variables in the data set.

3. Is there a possible pairwise linear relationship among the quantitative variables? Provide graphs and interpret the results.

B. Using the experiment data set experiment.csv:

Experiment: A study was conducted to determine the effect of 2 new feed formulation (1, 2) on the weight of eggs. Three species of ducks (A, B, C) were purposively selected for the study. The following data were generated.

Species-Feed	Weight	Species-Feed	Weight
A-1	5.6	B-2	7.3
A-1	5.8	B-2	7.1
A-2	6.1	C-1	6.3
A-2	6.3	C-1	6.2
B-1	8.1	C-2	6.8
B-1	8.2	C-2	6.9

1. Load the data into the object ds2.
   
2. Generate some descriptive summary statistics for all the variables in the data set.
   
3. Are the distributions of the weight similar across species?
   
4. Are the distributions of the weight similar across feed formulations?

---

EXERCISE 2 ANSWER KEY

A. Using the data set patients.csv:

1. Load the data into the object ds1.

#Kindly change the working directory where your patients.csv is located. THIS APPLIES TO ALL 'read.csv' and 'write.csv'
ds1 <- read.csv(file=("C:\\Users\\USER\\Desktop\\2019 Predictive Analytics Course in R\\R Scripts\\PA Course 1\\patients.csv"), header=TRUE)
ds1

2. Generate appropriate descriptive summary statistics for all the variables in the data set.

#Some Descriptive Statistics:
#Getting the sum of food expense
sum(ds1$foodexpense)

[1] 316658

#Getting the maximum of food expense
max(ds1$foodexpense)

[1] 9659

#Getting the range of age
range(ds1$age)

[1] 18 60

#Getting the correlation of weight and food expense
cor(ds1$weight,ds1$foodexpense)

[1] 0.2392273

#Getting the mean of weight
mean(ds1$weight)

[1] 65.86

#Getting the median of age
median(ds1$age)

[1] 41.5

#Install the package and call the library 'e1071'
library(e1071)
#Skewness (symmetry of distribution) of weight
skewness(ds1$weight)

[1] 0.04235242

#The skewness of weight is 0.04235242. It indicates that the eruption duration distribution is skewed towards the right.

#Graph the density plot to visualize the distribution
plot(density(ds1$weight))

3. Is there a possible pairwise linear relationship among the quantitative variables? Provide graphs and interpret the results.

#Getting the correlation of age and weight
cor(ds1$age,ds1$weight)

[1] -0.2155912

library("corrplot")
corr <- cor(ds1)
corr

                  patient          age     weight maritalstatus  foodexpense
patient        1.00000000  0.040889543  0.1569736   -0.04097957  0.150323967
age            0.04088954  1.000000000 -0.2155912   -0.19204346  0.006795935
weight         0.15697358 -0.215591218  1.0000000   -0.14157948  0.239227259
maritalstatus -0.04097957 -0.192043456 -0.1415795    1.00000000 -0.103970135
foodexpense    0.15032397  0.006795935  0.2392273   -0.10397013  1.000000000

#Sample Interpretation: age vs weight
# Correlation = -0.2155912. There is a weak negative linear relationship between age and weight.

#Seven different visualization methods can be used : “circle”, “square”, “ellipse”, “number”, “shade”, “color”, “pie”.
#I used the method: color
corrplot(corr, method = c("color"))

B. Using the experiment data set experiment.csv:

Experiment: A study was conducted to determine the effect of 2 new feed formulation (1, 2) on the weight of eggs. Three species of ducks (A, B, C) were purposively selected for the study. The following data were generated.

Species-Feed	Weight	Species-Feed	Weight
A-1	5.6	B-2	7.3
A-1	5.8	B-2	7.1
A-2	6.1	C-1	6.3
A-2	6.3	C-1	6.2
B-1	8.1	C-2	6.8
B-1	8.2	C-2	6.9

1. Load the data into the object ds2.

ds2 <- read.csv(file="C:\\Users\\USER\\Desktop\\2019 Predictive Analytics Course in R\\R Scripts\\PA Course 1\\experiment.csv")
ds2

2. Generate some descriptive summary statistics for all the variables in the data set.

#Some Descriptive Statistics:
#Getting the sum of weight
sum(ds2$weight)

[1] 80.7

#Getting the maximum of weight
max(ds2$weight)

[1] 8.2

#Getting the range of weight
range(ds2$weight)

[1] 5.6 8.2

#Getting the mean of weight
mean(ds2$weight)

[1] 6.725

#Getting the median of weight
median(ds2$weight)

[1] 6.55

#Install the package and call the library 'e1071'
library(e1071)
#Skewness (symmetry of distribution) of weight
skewness(ds2$weight)

[1] 0.4674504

#The skewness of weight is 0.04235242. It indicates that the eruption duration distribution is skewed towards the right.

#Graph the density plot to visualize the distribution in terms of weight
plot(density(ds2$weight))

3. Are the distributions of the weight similar across species?

#Install the package and call the library 'lattice'
library(lattice)
#Graph the density plot to visualize the distribution of the different species in terms of weight
densityplot(~weight, group = ds2$species, data = ds2, auto.key = TRUE)

4. Are the distributions of the weight similar across feed formulations?

#Install the package and call the library 'lattice'
library(lattice)
#Graph the density plot to visualize the distribution of the different feed formulation in terms of weight
densityplot(~weight, group = ds2$feed, data = ds2, auto.key = TRUE)