# My Course Project 2 of the Exploratory Data Analysis by Johns Hopkins University on Coursera.
# You can find all the data here.

# The overall goal of this assignment is to explore the National Emissions Inventory database and see what it say
# about fine particulate matter pollution in the United states over the 10-year period 1999–2008. You may use
# any R package you want to support your analysis. For each question/task you will need to make a single plot.
# Unless specified, you can use any plotting system in R to make your plot.


# QUESTION 1: Have total emissions from PM2.5 decreased in the United States from 1999 to 2008?
# Using the base plotting system, make a plot showing the total PM2.5 emission from all sources for each of the
# years 1999, 2002, 2005, and 2008.

# 1. Read and prepare data
NEI <- readRDS("summarySCC_PM25.rds")
# I use dplyr
library(dplyr)
NEIdp <- tbl_df(NEI)
# table with emissions per year
EmissionsForYear = summarize(group_by(NEIdp, year), sum(Emissions))
# renamed columns
colnames(EmissionsForYear) <- c("Year", "Emissions")
# new column with emissions in million (for the y axis)
EmissionsForYear$EmissionsInMillions = EmissionsForYear$Emissions / 1000000

# 2. Plot (png file)
png('plot1.png')
barplot(EmissionsForYear$EmissionsInMillions, names.arg=EmissionsForYear$Year, col="blue", xlab='Years', ylab='Emissions (PM 2.5) in millions', main =  'Emissions (PM 2.5) per year')
dev.off()

# QUESTION 2: Have total emissions from PM2.5 decreased in the Baltimore City, Maryland (fips == "24510")
# from 1999 to 2008? 
# Use the base plotting system to make a plot answering this question.

# 1. Read and prepare data
NEI <- readRDS("summarySCC_PM25.rds")
library(dplyr)
NEIdp <- tbl_df(NEI)
# table with emissions per year of Baltimore
BaltimoraEmissionsForYear = summarize(group_by(filter(NEIdp, fips=="24510"), year), sum(Emissions))
# renamed columns
colnames(BaltimoraEmissionsForYear) <- c("Year", "Emissions")

# 2. Plot
png('plot2.png')
barplot(BaltimoraEmissionsForYear$Emissions, names.arg=EmissionsForYear$Year, col="red", xlab='Years', ylab='Emissions (PM 2.5)', main =  'Baltimore City: Emissions (PM 2.5) per year')
dev.off()

# QUESTION 3: Of the four types of sources indicated by the type (point, nonpoint, onroad, nonroad) variable,
# which of these four sources have seen decreases in emissions from 1999-2008 for Baltimore City?
# Which have seen increases in emissions from 1999-2008?
# Use the ggplot2 plotting system to make a plot answer this question.

# 1. Read and prepare data
NEI <- readRDS("summarySCC_PM25.rds")
library(dplyr)
NEIdp <- tbl_df(NEI)
# Baltimore emissions per year and type
BaltimoraEmissionsTypeYears = summarize(group_by(filter(NEIdp, fips=="24510"),type,year),sum(Emissions))
# renamed columns
colnames(BaltimoraEmissionsTypeYears) <- c("Type","Year", "Emissions")
# year converted to string
BaltimoraEmissionsTypeYears$Year <- as.character(BaltimoraEmissionsTypeYears$Year)

# 2. Plot
library(ggplot2)
png('plot3.png')
qplot(Year,data=BaltimoraEmissionsTypeYears, geom="bar", weight=Emissions, facets=.~Type, fill=Year, main='Baltimore City: Emissions (PM 2.5) per year and type', xlab='', ylab = 'Emissions (PM 2.5)')
dev.off()

# QUESTION 4: Across the United States, how have emissions from coal combustion-related sources changed
# from 1999–2008?

# 1. Read and prepare data
NEI <- readRDS("summarySCC_PM25.rds")
SCC <- readRDS("Source_Classification_Code.rds")
# merge the tables
NEISCC <- merge(NEI, SCC, by="SCC")
library(dplyr)
NEISCCdp <- tbl_df(NEISCC)
# search for "coal" and create boolean column
NEISCCdp <- mutate(NEISCCdp, coal = grepl("coal", NEISCCdp$Short.Name, ignore.case=TRUE)) 
#renamed columns
EmissionsCoalYear <- summarize(group_by(filter(NEISCCdp, coal==TRUE),year),sum(Emissions))
colnames(EmissionsCoalYear) <- c("Year", "Emissions") 
# year converted to string
EmissionsCoalYear$Year <- as.character(EmissionsCoalYear$Year)
 #new column with emissions in thousands (for the y axis)
EmissionsCoalYear$EmissionsInTousands = EmissionsCoalYear$Emissions/1000

# 2. Plot
library(ggplot2)
png('plot4.png')
g <- ggplot(EmissionsCoalYear, aes(Year, EmissionsInTousands))
g+geom_bar(stat='identity')+labs(title="Emissions from coal combustion-related sources", x="Years",y="Emissions (PM 2.5) in thousands")
dev.off()

# QUESTION 5: How have emissions from motor vehicle sources changed from 1999-2008 in Baltimore City?

# 1. Read and prepare data
NEI <- readRDS("summarySCC_PM25.rds")
library(dplyr)
NEIdp <- tbl_df(NEI)
 # table with Motor Vehicle emissions per year of Baltimore
BaltimoraEmissionsVehicle = summarize(group_by(filter(NEIdp, fips=="24510", type=='ON-ROAD'), year), sum(Emissions)) 
#renamed columns
colnames(BaltimoraEmissionsVehicle) <- c("Year", "Emissions")

# 2. Plot
png('plot5.png')
g <- ggplot(BaltimoraEmissionsVehicle, aes(Year,Emissions))+scale_x_continuous(breaks = c(1999,2002,2005,2008))
g+geom_point(size=4, color='orange')+geom_line(size=1.5,color='orange')+labs(title="Baltimore City: Emissions of motor vehicle", x="Years",y="Emissions (PM 2.5)")
dev.off()

# QUESTION 6: Compare emissions from motor vehicle sources in Baltimore City with emissions from motor
# vehicle sources in Los Angeles County, California (fips == "06037"). Which city has seen greater changes over
# time in motor vehicle emissions?

# 1. Read and prepare data
NEI <- readRDS("summarySCC_PM25.rds")
library(dplyr)
NEIdp <- tbl_df(NEI)
# table with Motor Vehicle emissions per year of Baltimore
BaltimoraEmissionsVehicle = summarize(group_by(filter(NEIdp, fips=='24510', type=='ON-ROAD'), year), sum(Emissions)) 
# add place name
BaltimoraEmissionsVehicle = mutate(BaltimoraEmissionsVehicle, Place = 'Baltimore City')
# table with Motor Vehicle emissions per year of LA County
LAEmissionsVehicle = summarize(group_by(filter(NEIdp, fips=='06037', type=='ON-ROAD'), year), sum(Emissions))
# add place name
LAEmissionsVehicle = mutate(LAEmissionsVehicle, Place = 'Los Angeles County')
# union of the tables
BaltimoreLA <- rbind(BaltimoraEmissionsVehicle,LAEmissionsVehicle)
#renamed columns
colnames(BaltimoreLA) <- c('Year', 'Emissions', 'Place')
# year converted to string
BaltimoreLA$Year <- as.character(BaltimoreLA$Year)

# 2. Plot
png('plot6.png')
qplot(Year,data=BaltimoreLA, geom="bar", weight=Emissions, facets=.~Place, fill=Year, main='Baltimore City/Los Angeles County: Emissions of motor vehicle', xlab='', ylab = 'Emissions (PM 2.5)')
dev.off()