## Week 6 Homework

## Step 1: Load the data
install.packages("ggplot2")
install.packages("reshape2") 
library(ggplot2)
library(reshape2)

air <- airquality
View(air)
str(air)

## Step 2: Clean the data
# Check for NA's in the columns
any(is.na(air$Ozone)) # do any NA's exist in Ozone col? returns T or F

length(air[air=='NA'])
colnames(air)[colSums(is.na(air))>0]

# Replace NA's with mean of column
air$Ozone[is.na(air$Ozone)] <- mean(air$Ozone,na.rm=TRUE) # replace NA's in Ozone col with mean of col (where NA is discarded when calculating the mean)
air$Solar.R[is.na(air$Solar.R)] <- mean(air$Solar.R,na.rm=TRUE) # replace NA's in Solar.R col with mean of col
View(air)

## Step 3: Understand the data distribution
# histograms for each of the variables
# histogram for ozone
g.ozone <- ggplot(air, aes(x=Ozone)) + geom_histogram(binwidth=5, color="black", fill="white")
g.ozone # inspect histogram

# histogram for solar
g.solar <- ggplot(air, aes(x=Solar.R)) + geom_histogram(binwidth=5, color="black", fill="white")
g.solar # inspect histogram

# histogram for wind
g.wind <- ggplot(air, aes(x=Wind)) + geom_histogram(binwidth=5, color="black", fill="white")
g.wind # inspect histogram

# histogram for temp
g.temp <- ggplot(air, aes(x=Temp)) + geom_histogram(binwidth=5, color="black", fill="white")
g.temp # inspect histogram

# histogram for month
g.month <- ggplot(air, aes(x=Month)) + geom_histogram(binwidth=1, color="black", fill="white")
g.month # inspect histogram

# histogram for day
g.day <- ggplot(air, aes(x=Day)) + geom_histogram(binwidth=5, color="black", fill="white")
g.day # inspect histogram

# boxplot for ozone
gbox.ozone <- ggplot(air, aes(x=factor(0),Ozone)) + geom_boxplot()
gbox.ozone # inspect boxplot

# boxplot for wind
gbox.wind <- ggplot(air, aes(x=factor(0),Solar.R)) + geom_boxplot()
gbox.wind # inspect boxplot

## Step 3: Explore how the data changes over time
# Paste value in Month and Day columns and 1973 and assign to column "Date"
air$Date <- as.Date(paste("1973", air$Month, air$Day, sep="-"),"%Y-%m-%d")
str(air) # check structure to see if conversion works
# create line charts for ozone, temp, wind and Solar.R
# line chart for ozone
gline.ozone <- ggplot(air, aes(x=Date, y=Ozone)) + geom_line()
gline.ozone

# line chart for temp
gline.temp <- ggplot(air, aes(x=Date, y=Temp)) + geom_line()
gline.temp

# line chart for wind
gline.wind <- ggplot(air, aes(x=Date, y=Wind)) + geom_line()
gline.wind

# line chart for solar
gline.solar <- ggplot(air, aes(x=Date, y=Solar.R)) + geom_line()
gline.solar

# create one chart with 4 lines
air <- air[,-5:-6]

airLong <- melt(air, id="Date")
# create chart with all 4 variables
gline.all <- ggplot(airLong, aes(x=Date, y=value, color=variable)) + geom_line() + geom_smooth()
gline.all # inspect chart

#Another way to do it, create line chart for all 4 elements in one
ggplot(air, aes(x=Date)) + 
  geom_line(aes(y=Ozone, color="Ozone")) + 
  geom_line(aes(y=Temp, color="Temp")) +  
  geom_line(aes(y=Wind, color="Wind")) +
  geom_line(aes(y=Solar.R, color="Solar.R")) +  theme(plot.title=element_text(hjust=.5)) + 
  labs(title="1973 Environment", x="Date", y="Ozone / Temp / Wind / Solar.R")

## Step 4: Look at all the data via a Heatmap
htmap <- ggplot(airLong, aes(x=Date, y=variable)) + geom_tile(aes(fill=value)) + scale_fill_gradient(low="white", high="red")
htmap # inspect heatmap

## Step 5: Look at all the data via a scatter chart
gscatter <- ggplot(air, aes(x=Wind, y=Temp)) + geom_point(aes(size=Ozone, color=Solar.R))
gscatter # inspect scatter plot


## Step 6: Final Analysis