R Week 4 - Data Analysis - Water Hardness and Mortality

Load libraries and retreive data from remote URL

# Load libraries RCurl -> for loading data from URL, plotly -> for making plots interactive, ggrepel -> for Text Labels
library(RCurl)
## Loading required package: bitops
library(plotly)
## Loading required package: ggplot2
## 
## Attaching package: 'plotly'
## The following object is masked from 'package:ggplot2':
## 
##     last_plot
## The following object is masked from 'package:graphics':
## 
##     layout
library(ggrepel)

# Load Data from Remote URL
rData <- getURL('http://vincentarelbundock.github.io/Rdatasets/csv/HSAUR/water.csv')

# Read CSV data in lData (local data), test connection , header by default is true
lData <- read.csv(textConnection(rData))

# View Data
knitr::kable(lData)
X location town mortality hardness
1 South Bath 1247 105
2 North Birkenhead 1668 17
3 South Birmingham 1466 5
4 North Blackburn 1800 14
5 North Blackpool 1609 18
6 North Bolton 1558 10
7 North Bootle 1807 15
8 South Bournemouth 1299 78
9 North Bradford 1637 10
10 South Brighton 1359 84
11 South Bristol 1392 73
12 North Burnley 1755 12
13 South Cardiff 1519 21
14 South Coventry 1307 78
15 South Croydon 1254 96
16 North Darlington 1491 20
17 North Derby 1555 39
18 North Doncaster 1428 39
19 South East Ham 1318 122
20 South Exeter 1260 21
21 North Gateshead 1723 44
22 North Grimsby 1379 94
23 North Halifax 1742 8
24 North Huddersfield 1574 9
25 North Hull 1569 91
26 South Ipswich 1096 138
27 North Leeds 1591 16
28 South Leicester 1402 37
29 North Liverpool 1772 15
30 North Manchester 1828 8
31 North Middlesbrough 1704 26
32 North Newcastle 1702 44
33 South Newport 1581 14
34 South Northampton 1309 59
35 South Norwich 1259 133
36 North Nottingham 1427 27
37 North Oldham 1724 6
38 South Oxford 1175 107
39 South Plymouth 1486 5
40 South Portsmouth 1456 90
41 North Preston 1696 6
42 South Reading 1236 101
43 North Rochdale 1711 13
44 North Rotherham 1444 14
45 North St Helens 1591 49
46 North Salford 1987 8
47 North Sheffield 1495 14
48 South Southampton 1369 68
49 South Southend 1257 50
50 North Southport 1587 75
51 North South Shields 1713 71
52 North Stockport 1557 13
53 North Stoke 1640 57
54 North Sunderland 1709 71
55 South Swansea 1625 13
56 North Wallasey 1625 20
57 South Walsall 1527 60
58 South West Bromwich 1627 53
59 South West Ham 1486 122
60 South Wolverhampton 1485 81
61 North York 1378 71 
# Data Summary
summary(lData)
##        X       location          town      mortality       hardness     
##  Min.   : 1   North:35   Bath      : 1   Min.   :1096   Min.   :  5.00  
##  1st Qu.:16   South:26   Birkenhead: 1   1st Qu.:1379   1st Qu.: 14.00  
##  Median :31              Birmingham: 1   Median :1555   Median : 39.00  
##  Mean   :31              Blackburn : 1   Mean   :1524   Mean   : 47.18  
##  3rd Qu.:46              Blackpool : 1   3rd Qu.:1668   3rd Qu.: 75.00  
##  Max.   :61              Bolton    : 1   Max.   :1987   Max.   :138.00  
##                          (Other)   :55

Create Scatter Plot

# Scatter Diagram 1
b <- ggplot(lData, aes(x = mortality , y = hardness)) 
b + geom_point(aes(color = hardness, size = mortality, shape = factor(location))) +  scale_colour_gradient(low = "purple")

# Lets make it interactive
ggplotly()

In the above scatter plot we can see the clear distinction between triangles and circle. The traingles have higher Hardness rate and the Circles have a higher mortality rate.

Create North South Scatter Plot

# Scatter Diagram 2
ggplot(subset(lData, location %in% c("North", "South")),
       aes(x=mortality, y=hardness, color=location)) + geom_point() +  geom_smooth() + geom_text_repel(aes(label=town), size = 3)

In the above scatter plot we can very well see that the North has a higher mortality rate than the South. We can also see that the twon Ipswich has the lowest mortality rate (1096) and the town Salford has the highest mortality rate of 1987.

Mortality Histogram

# Create Histogram for Mortality
ggplot(data=lData, aes(lData$mortality)) +
  geom_histogram(aes(y =..density.., fill=..count..), breaks=seq(1000, 2000, by = 100), alpha = .5 ) +
  geom_density(col=1) + labs(title="Histogram for Mortality", x="Mortality", y="Count")  +
  geom_vline(aes(xintercept=mean(mortality)), color="blue", linetype="dashed", size=1)

The above histogram shows that the maximum frequency of Mortality Rate is between 1300 to 1600. the dashed blue line represent Mean of the distribution.

Water Hardness Histogram

ggplot(data=lData, aes(lData$hardness)) +
  geom_histogram(aes(y =..density.., fill=..count..), breaks=seq(0, 150, by = 10), alpha = .5 ) +
  geom_density(col=2) + labs(title="Histogram for Water Hardness", x="Water Hardness", y="Count") +
  geom_vline(aes(xintercept=mean(hardness)), color="blue", linetype="dashed", size=1)

As we can see in the histogram, the frequncy of water hardness between 10 to 20 is maximum and the dashed blue line represent Mean of the distribution.

Mortality Rate comparison between North and South Location

# North vs South Mortality Histogram
ggplot(lData, aes(x=mortality, color=location, fill=location)) +
  geom_histogram(position="identity",breaks=seq(1000, 2000, by = 100), alpha=0.5)+
  scale_color_manual(values=c("#999999", "#E69F00", "#56B4E9"))+
  scale_fill_manual(values=c("#999999", "#E69F00", "#56B4E9"))+
  labs(title="Mortality histogram plot",x="Mortality", y = "Count")

# Lets make it interactive
ggplotly()

The Mortality Rate for South Region lies between 1000 to 1700 whereas the Mortslity Rate of North Region lies between 1300 to 2000.

Water Hardness comparison between North and South Location

# North vs South Water Hardness Histogram
ggplot(lData, aes(x=hardness, color=location, fill=location)) +
  geom_histogram(position="identity",breaks=seq(0, 150, by = 10), alpha=0.5)+
  scale_color_manual(values=c("#999999", "#BEE181", "#5B6BAA"))+
  scale_fill_manual(values=c("#999999", "#BEE181", "#5B6BAA"))+
  labs(title="Water Hardness histogram plot",x="hardness", y = "Count")

# Lets make it interactive
ggplotly()

The North region has highest density of hardness water rate with low value whereas the South region has Water Hardness rate with high value.

Bar Plot Hardness vs Mortality

ggplot(lData, aes(x=mortality, y=hardness))  + geom_bar(stat = "identity") + geom_smooth()

The above bar plot reveals that as the Water Hardness rate is inversely proportional to Mortality Rate i.e as hardness decreases, mortality increases.

Mortality Rate North vs South

# Dotplot: Grouped Sorted and Colored
# Sort by mortality, group and color by location
xm <- lData[order(lData$mortality),] # sort by mortality
xm$loc <- factor(xm$location) # it must be a factor
xm$color[xm$location=="North"] <- "red"
xm$color[xm$location=="South"] <- "blue"
dotchart(xm$mortality,labels=row.names(xm),cex=.50,groups= xm$location,
         main="Mortality rates grouped by location",
         xlab="Mortality rate", gcolor="green", color=xm$color)

Water Hardness North vs South

# Dotplot: Grouped Sorted and Colored
# Sort by hardness, group and color by location
wHard <- lData[order(lData$hardness),] # sort by hardness
wHard$loc <- factor(wHard$location) # it must be a factor
wHard$color[wHard$location=="North"] <- "red"
wHard$color[wHard$location=="South"] <- "blue"
dotchart(wHard$hardness,labels=row.names(xm),cex=.50,groups= wHard$location,
         main="Water Hardness grouped by location",
         xlab="Water Hardness", gcolor="Gray", color=wHard$color)

Mortality Rate Boxplot based on Location

#Mortality Boxplot based on Location
boxplot(lData$mortality~lData$location,data=lData, main="Mortality data", 
        xlab="Location", ylab="Mortality data")

The boxplot clearly shows Mortality Rate is comparatively higher in North Region when compared with South. Observe the Median of the both the Region.

The middle 50% of the Mortality Rate in North Region lies between 1550 to 1700, whereas the median of the data of North Region is little higher than 1600. As observed, we can see only one data point as outlier.

The median of the moertality rate in South Region is littler higher than 1300. The middle 50% lies between 1250 to 1500.

Water Hardness Rate Boxplot based on Location

# Water Hardness based on Location
boxplot(lData$hardness~lData$location,data=lData, main="Water Hardness", 
        xlab="Location", ylab="Water Hardness")

Water Hardness in the North region is skewed right i.e most of the values are greater than median, whereas Water Hardness in the South region is skewed left i.e most of the values are lower than the median.

Water Hardness Rate vs Mortality Rate Boxplot based on Location

#  Hardness vs Mortality Boxplot based on Location
ggplot(lData, aes(x=mortality, y=hardness, color=location)) + geom_boxplot()
## Warning: position_dodge requires non-overlapping x intervals