Análisis y visualización de consumo eléctrico

Se analiza el consumo eléctrico de una casa con los datos tomados a lo largo de 4 años. En la casa hay instalados 3 submedidores, para distintas zonas de estas. Analizamos las medidas tomadas por éstos el y el consumo global activo.

Cargamos librerías usadas y los datos a analizar. Juntamos fecha y hora en una columna nueva

library(dplyr)
library(tidyr)
library(ggplot2)
library(lubridate)
#detach(package:plyr)
setwd("C://Users/Pau A/Documents/Data Analysis Course/3- Deep Analytics and Visualization/Task 1")

The working directory was changed to C:/Users/Pau A/Documents/Data Analysis Course/3- Deep Analytics and Visualization/Task 1 inside a notebook chunk. The working directory will be reset when the chunk is finished running. Use the knitr root.dir option in the setup chunk to change the the working directory for notebook chunks.

Consumption <- read.csv("household_power_consumption.txt", TRUE, ";", na.strings = c("NA","?"))

Preprocesamos los datos. Convertimos a númerico las columnas referentes al consumo global activo, reactivo, al voltaje y a los submedidores. Creamos nuevas columnas con los días de la semana y el mes. Y finalmente creamos un dataset sin NAs.

### Pre-proceso de los datos. 
#Creamos una nueva columna con Fecha y Hora
Consumption <-cbind(Consumption,paste(Consumption$Date,Consumption$Time), stringsAsFactors=FALSE)
colnames(Consumption)[10] <-"DateTime"
Consumption <- Consumption[,c(ncol(Consumption), 1:(ncol(Consumption)-1))]
#Añadimos una columna con el GAP a W·h
Consumption <- cbind(Consumption, Consumption$Global_active_power*1000/60)
colnames(Consumption)[11] <- "Global_Consumption"
#Pasamos a númerico
Consumption$Global_active_power <- as.numeric(Consumption$Global_active_power)
Consumption$Global_reactive_power <- as.numeric(Consumption$Global_reactive_power)
Consumption$Voltage <- as.numeric(Consumption$Voltage)
Consumption$Sub_metering_1 <- as.numeric(Consumption$Sub_metering_1)
Consumption$Sub_metering_2 <- as.numeric(Consumption$Sub_metering_2)
Consumption$Sub_metering_3 <- as.numeric(Consumption$Sub_metering_3)
Consumption$Global_Consumption <- as.numeric(Consumption$Global_Consumption)
#Creamos una nueva columna con los d??as de la semana
Consumption <- cbind(Consumption,weekdays(as.Date(Consumption$Date, '%d/%m/%Y')), stringsAsFactors=FALSE)
colnames(Consumption)[12] <- "Day"
#Creamos una nueva columna con el mes
Consumption <- cbind(Consumption,months(as.Date(Consumption$Date, '%d/%m/%Y')), stringsAsFactors=FALSE)
colnames(Consumption)[13] <- "Month"
#Creamos una columna de mes/año
MonthYear <- separate(Consumption, Date, into=c("day", "month", "year"))
MonthYear <- paste(MonthYear$month, MonthYear$year, sep="/")
Consumption <-cbind(Consumption, MonthYear)
#Filtramos el dataset en datos de cada 15 minutos
Minute <- separate(Consumption, Time, into = c("Hour", "Minute", "Second"))
Minute <- Minute$Minute
Consumption <-cbind(Consumption, Minute)
Consumption$Minute <- as.numeric(Consumption$Minute)
Reduced_data <- subset(Consumption, Minute==1 | Minute==31)
#Añadimos columna de semanas 
semana <- c(1:205) 
initial <- c()
for (i in semana){
  initial <- c(initial, rep(semana[[i]], 10080))
}
Consumption$Week <- c(initial, rep(206, 8859))
#Dataset sin missing values
cleanConsumption <- na.omit(Consumption)

Analizamos el consumo en los meses de Enero y Agosto, comparando de Lunes a Viernes y los fines de semana, para los submedidores y el consumo global activo.

###Enero
january <- subset(Reduced_data, Month == "enero")
january_weekend <- subset(Reduced_data, Month == "enero" & (Day=="s?bado" | Day=="domingo"))
january_weekday <- subset(Reduced_data, Month == "enero" & Day!="s?bado" & Day!="domingo")
#Sub1
j_sub_1_wd <- ggplot(january_weekday, aes(x=Time)) + geom_point(aes(y=Sub_metering_1), color = "red")
j_sub_1_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 1") + ggtitle("Sub-Metering 1 en Enero, Lunes-Viernes")

j_sub_1_we <- ggplot(january_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_1), color = "blue")
j_sub_1_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 1") + ggtitle("Sub-Metering 1 en Enero, fin de semana")

#Sub2
j_sub_2_wd <- ggplot(january_weekday, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "red")
j_sub_2_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 2") + ggtitle("Sub-Metering 2 en Enero, Lunes-Viernes")

j_sub_2_we <- ggplot(january_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "blue")
j_sub_2_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 2") + ggtitle("Sub-Metering 2 en Enero, fin de semana")

#Sub3
j_sub_3_wd <- ggplot(january_weekday, aes(x=Time)) + geom_point(aes(y=Sub_metering_3), color = "red")
j_sub_3_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 3") + ggtitle("Sub-Metering 3 en Enero, Lunes-Viernes")

j_sub_3_we <- ggplot(january_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "blue")
j_sub_3_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 3") + ggtitle("Sub-Metering 3 en Enero, fin de semana")

#GAP
j_gap_wd <- ggplot(january_weekday, aes(x=Time)) + geom_point(aes(y=Global_Consumption),color = "red")
j_gap_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Consumo Global Activo") + ggtitle("Consumo Global Activo en Enero, Lunes-Viernes")

j_gap_we <- ggplot(january_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "blue")
j_gap_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Consumo Global Activo") + ggtitle("Consumo Global Activo en Enero, fin de semana")

###Agosto
august <- subset(Reduced_data, Month == "agosto")
august_weekend <- subset(Reduced_data, Month == "agosto" & (Day=="s?bado" | Day=="domingo"))
august_weekday <- subset(Reduced_data, Month == "agosto" & Day!="s?bado" & Day!="domingo")
#Sub1
a_sub_1_wd <- ggplot(august_weekday, aes(x=Time)) + geom_point(aes(y=Sub_metering_1), color = "red")
a_sub_1_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 1") + ggtitle("Sub-Metering 1 en Agosto, Lunes-Viernes")

a_sub_1_we <- ggplot(august_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_1), color = "blue")
a_sub_1_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 1") + ggtitle("Sub-Metering 1 en Agosto, fin de semana")

#Sub2
a_sub_2_wd <- ggplot(august_weekday, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "red")
a_sub_2_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 2") + ggtitle("Sub-Metering 2 en Agosto, Lunes-Viernes")

a_sub_2_we <- ggplot(august_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "blue")
a_sub_2_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 2") + ggtitle("Sub-Metering 2 en Agosto, fin de semana")

#Sub3
a_sub_3_wd <- ggplot(august_weekday, aes(x=Time)) + geom_point(aes(y=Sub_metering_3), color = "red")
a_sub_3_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 3") + ggtitle("Sub-Metering 3 en Agosto, Lunes-Viernes")

a_sub_3_we <- ggplot(august_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "blue")
a_sub_3_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Sub-Metering 3") + ggtitle("Sub-Metering 3 en Agosto, fin de semana")

#GAP
a_gap_wd <- ggplot(august_weekday, aes(x=Time)) + geom_point(aes(y=Global_Consumption),color = "red")
a_gap_wd + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Consumo Global Activo") + ggtitle("Consumo Global Activo en Agosto, Lunes-Viernes")

a_gap_we <- ggplot(august_weekend, aes(x=Time)) + geom_point(aes(y=Sub_metering_2), color = "blue")
a_gap_we + theme(axis.text.x = element_text(size=6, angle=90)) + xlab("Hora") + ylab("Consumo Global Activo") + ggtitle("Consumo Global Activo en Agosto, fin de semana")

Realizamos un análisis de las medias en cada hora para los meses de Enero y Agosto.

###Calculamos las medias por cada hora en cada mes
#Enero
january_mean <- january %>%
  group_by(Time) %>%
  summarise(
    Sub_1_mean_january <- mean(Sub_metering_1, na.rm = TRUE),
    Sub_2_mean_january <- mean(Sub_metering_2, na.rm = TRUE),
    Sub_3_mean_january <- mean(Sub_metering_3, na.rm = TRUE),
    GAP_mean_january <- mean(Global_Consumption, na.rm = TRUE)
  )
colnames(january_mean) <- c("Time","Sub_1_mean","Sub_2_mean","Sub_3_mean","GAP_mean")
j_mean <- ggplot(january_mean, aes(x=Time, group=1)) + geom_line(aes(y=Sub_1_mean, color="Sub-Metering 1")) + 
  geom_line(aes(y=Sub_2_mean, color="Sub-Metering 2")) + 
  geom_line(aes(y=Sub_3_mean, color="Sub-Metering 3")) + 
  geom_line(aes(y=GAP_mean, color="Global Active Power")) +
  scale_color_manual("", breaks=c("Sub-Metering 1","Sub-Metering 2","Sub-Metering 3","Global Active Power"),
                     values=c("red","blue","brown","green"))
j_mean + theme(axis.text.x = element_text(size=6, angle=90)) + ylab("W?h") +
  labs(title="Consumo medio en un día de Enero") + xlab("Hora")

#Agosto
august_mean <- august %>%
  group_by(Time) %>%
  summarise(
    Sub_1_mean_august <- mean(Sub_metering_1, na.rm = TRUE),
    Sub_2_mean_august <- mean(Sub_metering_2, na.rm = TRUE),
    Sub_3_mean_august <- mean(Sub_metering_3, na.rm = TRUE),
    GAP_mean_august <- mean(Global_Consumption, na.rm = TRUE)
  )
colnames(august_mean) <- c("Time","Sub_1_mean","Sub_2_mean","Sub_3_mean","GAP_mean")
a_mean <- ggplot(august_mean, aes(x=Time, group=1)) + geom_line(aes(y=Sub_1_mean, color="Sub-Metering 1")) + 
             geom_line(aes(y=Sub_2_mean, color="Sub-Metering 2")) + 
            geom_line(aes(y=Sub_3_mean, color="Sub-Metering 3")) + 
            geom_line(aes(y=GAP_mean, color="Global Active Power")) +
            scale_color_manual("", breaks=c("Sub-Metering 1","Sub-Metering 2","Sub-Metering 3","Global Active Power"),
                                  values=c("red","blue","brown","green"))
a_mean + theme(axis.text.x = element_text(size=6, angle=90)) + ylab("W?h") +
                labs(title="Consumo medio en un día de Agosto") + xlab("Hora")

Posterioremente realizamos una visualización del consumo a lo largo del año 2009:

Year <- separate(cleanConsumption, Date, into = c("day", "month", "Year"))
Date_2 <- as.Date(paste(Year$day, Year$month, Year$Year), "%d %m %Y", tz="GMT")
Year <- Year$Year
cleanConsumption <-cbind(cleanConsumption, Year, Date_2)
cleanConsumption$Year <- as.numeric(cleanConsumption$Year)
data_2009 <- subset(cleanConsumption, Year==4)
year <- data_2009 %>%
  group_by(Date_2) %>%
  summarise(
    Sub_1_mean <- mean(Sub_metering_1, na.rm = TRUE),
    Sub_2_mean <- mean(Sub_metering_2, na.rm = TRUE),
    Sub_3_mean <- mean(Sub_metering_3, na.rm = TRUE),
    GAP_mean <- mean(Global_Consumption, na.rm = TRUE)
  )
colnames(year) <- c("Date","Sub_1_mean","Sub_2_mean","Sub_3_mean","GAP_mean")
consumption_2009 <- ggplot(year, aes(x=Date, group=1)) + geom_line(aes(y=Sub_1_mean, color="Sub-Metering 1")) + 
  geom_line(aes(y=Sub_2_mean, color="Sub-Metering 2")) + 
  geom_line(aes(y=Sub_3_mean, color="Sub-Metering 3")) + 
  geom_line(aes(y=GAP_mean, color="Global Active Power")) +
  scale_color_manual("", breaks=c("Sub-Metering 1","Sub-Metering 2","Sub-Metering 3","Global Active Power"),
                     values=c("red","blue","brown","green"))
consumption_2009 + theme(axis.text.x = element_text(size= 10, angle=45)) + ylab("W·h") +
  labs(title="Consumo en 2009")

Y finalmente reealizamos un análisis del consumo a lo largo de todos los años. Podemos observar un comportamiento similar al sinusoidal con subidas de consumo en invierno y bajadas en verano.

#Evolución del consumo a lo largo de todos los años
Year <- separate(cleanConsumption, Date, into = c("day", "month", "Year"))
Date_2 <- as.Date(paste(Year$day, Year$month, Year$Year), "%d %m %Y", tz="GMT")
day <- Year$day
Year <- Year$Year
cleanConsumption <-cbind(cleanConsumption, Year, Date_2, day)
lesspoints <- subset(cleanConsumption, day==1 | day==5 | day==10 | day==15 | day==20 | day==25 | day==30)
evolution <- lesspoints %>%
  group_by(Date_2) %>%
  summarise(
    Sub_1_mean <- mean(Sub_metering_1, na.rm = TRUE),
    Sub_2_mean <- mean(Sub_metering_2, na.rm = TRUE),
    Sub_3_mean <- mean(Sub_metering_3, na.rm = TRUE),
    GAP_mean <- mean(Global_Consumption, na.rm = TRUE)
)
colnames(evolution) <- c("Date","Sub_1_mean","Sub_2_mean","Sub_3_mean","GAP_mean")
cons_evol <- ggplot(evolution, aes(x=Date, group=1)) + geom_line(aes(y=Sub_1_mean, color="Sub-Metering 1")) + 
  geom_line(aes(y=Sub_2_mean, color="Sub-Metering 2")) + 
  geom_line(aes(y=Sub_3_mean, color="Sub-Metering 3")) + 
  geom_line(aes(y=GAP_mean, color="Global Active Power")) +
  scale_color_manual("", breaks=c("Sub-Metering 1","Sub-Metering 2","Sub-Metering 3","Global Active Power"),
                     values=c("red","blue","brown","green"))
cons_evol + theme(axis.text.x = element_text(size= 10, angle=45)) + ylab("W?h") + xlab("Fecha") +
  labs(title="Evolución del consumo")