Regression start week3 part1
Sorting data
##################day time function
daytime<- function(data4=bikedata){
a<- strptime(data4$Start.Date,format="%d/%m/%Y %R",tz="UTC")
b<- format(a, "%R")
Time <- hour(hm(b))
timeofday <- hour(hm("00:00", "5:00", "12:00", "18:00", "23:59"))
names <- c("Night", "Morning", "Afternoon", "Evening")
z<-cut(x=Time, breaks=timeofday, labels=names, include.lowest=TRUE)
return(z)
}
#######################################################season
Season <- function(data2) {
d<- data2$Start.Date
WS <- as.Date("21/12/2017", format = "%d/%m/%Y") # Winter
SE <- as.Date("20/3/2017", format = "%d/%m/%Y") # Spring
SS <- as.Date("21/6/2017", format = "%d/%m/%Y") # Summer
FE <- as.Date("22/9/2017", format = "%d/%m/%Y") # Autumn
ifelse (d >= WS | d < SE, "Winter",
ifelse (d >= SE & d < SS, "Spring",
ifelse (d >= SS & d < FE, "Summer", "Autumn")))
}bikedata$Daytime<- daytime(data4=bikedata)
bikedata$Start.Date<- as.Date(strptime(bikedata$Start.Date,
format="%d/%m/%Y %H:%M",tz="UTC"))
bikedata$weekdays<- wday(bikedata$Start.Date,label = T)
bikedata$weekdays<- factor(bikedata$weekdays,ordered = FALSE)
bikedata$season<-as.factor(Season(data2=bikedata))
x<-plyr::count(bikedata,vars = c("Start.Date","Daytime","weekdays","season"))
colnames(x)[5]<- c("ntrips")
bikedata<-join(x,bikedata)## Joining by: Start.Date, Daytime, weekdays, seasonbikedata<- bikedata%>% dplyr::select("Start.Date","Daytime","weekdays","season","ntrips")
head(bikedata,10)## Start.Date Daytime weekdays season ntrips
## 1 2017-01-04 Night Wed Winter 292
## 2 2017-01-04 Night Wed Winter 292
## 3 2017-01-04 Night Wed Winter 292
## 4 2017-01-04 Night Wed Winter 292
## 5 2017-01-04 Night Wed Winter 292
## 6 2017-01-04 Night Wed Winter 292
## 7 2017-01-04 Night Wed Winter 292
## 8 2017-01-04 Night Wed Winter 292
## 9 2017-01-04 Night Wed Winter 292
## 10 2017-01-04 Night Wed Winter 292bikedata<- bikedata[!duplicated(bikedata),]
bikedata<- bikedata%>% dplyr::select("ntrips","weekdays","season","Daytime")
bikedata<- na.omit(bikedata)
head(bikedata,20)## ntrips weekdays season Daytime
## 1 292 Wed Winter Night
## 293 9487 Wed Winter Morning
## 9780 9366 Wed Winter Afternoon
## 19146 2810 Wed Winter Evening
## 21956 325 Thurs Winter Night
## 22281 10244 Thurs Winter Morning
## 32525 9752 Thurs Winter Afternoon
## 42277 3029 Thurs Winter Evening
## 45306 392 Fri Winter Night
## 45698 9950 Fri Winter Morning
## 55648 6701 Fri Winter Afternoon
## 62349 1767 Fri Winter Evening
## 64116 453 Sat Winter Night
## 64569 4230 Sat Winter Morning
## 68799 6711 Sat Winter Afternoon
## 75510 2328 Sat Winter Evening
## 77838 925 Sun Winter Night
## 78763 4166 Sun Winter Morning
## 82929 6491 Sun Winter Afternoon
## 89420 1753 Sun Winter Eveningstr(bikedata)## 'data.frame': 1148 obs. of 4 variables:
## $ ntrips : int 292 9487 9366 2810 325 10244 9752 3029 392 9950 ...
## $ weekdays: Factor w/ 7 levels "Sun","Mon","Tues",..: 4 4 4 4 5 5 5 5 6 6 ...
## $ season : Factor w/ 4 levels "Autumn","Spring",..: 4 4 4 4 4 4 4 4 4 4 ...
## $ Daytime : Factor w/ 4 levels "Night","Morning",..: 1 2 3 4 1 2 3 4 1 2 ...Exploratory analysis
###################weekdays and season
ggplot(data=bikedata,aes(x=weekdays,y=ntrips))+
geom_boxplot(aes(fill=season))+facet_grid(~season)+
ggtitle("Boxplot of the number of trips by weekdays and season")+
theme(plot.title = element_text(hjust = 0.5))+
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Relatively speaking, Weekends have fewer number of trips, especially during winter and autumn. However the seasonal patterns on its own does not seems to have a big influence.
##############weekdays and daytime
ggplot(data=bikedata,aes(x=weekdays,y=ntrips))+
geom_boxplot(aes(fill=Daytime))+facet_grid(~Daytime)+
ggtitle("Boxplot of the number of trips by weekdays and daytime")+
theme(plot.title = element_text(hjust = 0.5))+
theme(axis.text.x = element_text(angle = 45, hjust = 1))
A dramatically difference inbetween each daytimes which seems to be dominated by Morning and afternoon, suggesting citizens may cycle to and back from work. Change in weekdays also alter the daytime slightly, as there are fewer trips on morning in weekends, however during the night time, weekends have slightly higher number of trips relatively. These may leads to the conclusion that citizens tends to using bikes little more often at night time during weekends, and morning and afternoon time in weekdays.
#########Season and daytime
ggplot(data=bikedata,aes(x=Daytime,y=ntrips))+
geom_boxplot(aes(fill=season))+facet_grid(~season)+
ggtitle("Boxplot of the number of trips by season and daytime")+
theme(plot.title = element_text(hjust = 0.5))+
theme(axis.text.x = element_text(angle = 45, hjust = 1))
sum(bikedata[which(bikedata$season=="Summer"),]$ntrips)## [1] 2470355sum(bikedata[which(bikedata$season=="Autumn"),]$ntrips)## [1] 1497932sum(bikedata[which(bikedata$season=="Spring"),]$ntrips)## [1] 2906197sum(bikedata[which(bikedata$season=="Winter"),]$ntrips)## [1] 1560046seasonal changes have small impact on the daytime, in Evening and afternoon, the number of trips increases from autumn and dropped all the way down in Winter again.