.Rhistory

forecasts3
#on va avoir une period mean qui va nous donne forcast at eahc pouint h
#if we want to depict foreacts and fitted values along with time series
forecasts3 %>%
# Depicts the original time series and the forecasts
autoplot(manual_decomposition, level = FALSE) +
# Overlays the fitted values
geom_line(data = fitted_vals3, aes(y = .fitted), colour = "blue", linetype = "dashed")
library(fpp3)
## YOUR CODE GOES HERE
scs = scs_pedestrians%>%filter(Date>="2016-07-01"&Date<="2016-10-25")
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
scs_pedestrians
## YOUR CODE GOES HERE
scs_pedestrians %>% autoplot(Count) + scale_x_date(
breaks = "1 week",
minor_breaks = "5 week"
) + theme(axis.text.x = element_text(angle = 90))
scs_pedestrians$mean_val = mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians
## YOUR CODE GOES HERE
scs_pedestrians <-
scs_pedestrians %>%
mutate(`12-MA` = slider::slide_dbl(Count, mean,.before = 5, .after = 6, .complete =
TRUE),
`trend_class` = slider::slide_dbl(`12-MA`, mean,.before = 1, .after = 0, .complete =
TRUE))
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
# FEEDBACK: IF THE PERIOD IS ONE WEEK AND THE DATA IS DAILY, YOU NEED THE 7-MA
# NOT THE 2X12-MA. THIS CODE WAS COPIED FROM THE NOTES WITHOUT THINKING IF IT APPLIES AT ALL
scs_pedestrians <-
scs_pedestrians %>%
mutate(`12-MA` = slider::slide_dbl(Count, mean,.before = 5, .after = 6, .complete =
TRUE),
`trend_class` = slider::slide_dbl(`12-MA`, mean,.before = 1, .after = 0, .complete =
TRUE))
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
fit = scs_pedestrians %>%
model(
mean = MEAN(Count),
Drift = RW(Count ~ drift()),
SNaive = SNAIVE(Count)
)
fitted_vals <-
fit %>%
augment()
fitted_vals
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
fitted_vals <-
fit %>%
augment()
forecasts <- fit %>% forecast(h = 7)
forecasts
# Depict the forecasts
forecasts %>%
filter(.model == "SNaive")
autoplot(scs_pedestrians, level = FALSE) +
autolayer(fitted_vals, .fitted, colour = "blue", linetype = "dashed")
forecasts %>%
filter(.model == "mean")
autoplot(scs_pedestrians, level = FALSE) +
autolayer(fitted_vals, .fitted, colour = "blue", linetype = "dashed")
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
scs_pedestrians%>%ACF(Count)%>%autoplot()
fit2 = scs_pedestrians %>%
model(
mean = MEAN(Count),
)
fitted_vals2 <-
fit2 %>%
augment()
fitted_vals2%>%ACF(.innov)%>%autoplot()
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
fitted_vals %>%
filter(.model == "mean") %>%
autoplot(Count, colour = "gray") +
geom_line(aes(y=.fitted), colour = "blue", linetype = "dashed")
fit %>%
select(mean) %>%
gg_tsresiduals()
# Compute the mean of the residuals
xx = fitted_vals %>% as_tibble() %>%
filter(.model == "Naive") %>%
summarise(mean = mean(.innov, na.rm = TRUE))
xx <- filter(xx, .model=="mean")
library(fpp3)
library(fpp3)
## YOUR CODE GOES HERE
new_data <- pedestrian %>%
filter(Sensor=="Southern Cross Station" | Sensor=="Bourke Street Mall (North)") %>%
filter(Date_Time>="2016-07-01" & Date_Time<="2016-10-25") %>%
as_tibble() %>%
group_by(Sensor,Date) %>%
summarise(Total_Count=sum(Count))
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
new_data
## YOUR CODE GOES HERE
new_data %>%
ggplot(aes(x=Date, y=Total_Count,fill=Sensor))+
geom_line()+scale_x_date(breaks ="5 weeks", minor_breaks="1 weeks" )
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
new_data %>% group_by(Sensor) %>% summarise(Max=max(Total_Count))
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
scs_pedestrians
## YOUR CODE GOES HERE
scs_pedestrians %>% autoplot()+
scale_x_date(breaks ="5 weeks", minor_breaks="1 week" )
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians$mean_val = mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians
## YOUR CODE GOES HERE
manual_decompostion <- scs_pedestrians %>%
mutate(
`7-MA`=slider::slide_dbl(Count,mean,
.before=3,.after=3,.complete = TRUE )
) %>%
select(Sensor,Date,Count,mean_val,`7-MA`)
manual_decompostion %>%
autoplot(Count, colour="gray")+
geom_line(aes(y=`7-MA`),colour="red")+
geom_line(aes(y=mean_val ),colour="blue",linetype="dashed")
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
fit_mean <- scs_pedestrians %>%
model(mean= MEAN(Count))
fit_mean
fit_drift <- scs_pedestrians %>%
model(drift=RW(Count~ drift()))
fit_drift
fit_snaive <- scs_pedestrians %>%
model(snaive=SNAIVE(Count))
fit_snaive
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
fit_mean <- scs_pedestrians %>%
model(mean= MEAN(Count))
fit_mean
fit_drift <- scs_pedestrians %>%
model(drift=RW(Count~ drift()))
fit_drift
fit_snaive <- scs_pedestrians %>%
model(snaive=SNAIVE(Count))
fit_snaive
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
## YOUR CODE GOES HERE
#SNAIVE model
fitted_values_snaive <- fit_snaive %>% augment()
head(fitted_values_snaive)
forecast_snaive <- fit_snaive %>% forecast(h=7)
forecast_snaive %>% autoplot(scs_pedestrians,level=FALSE)+
autolayer(fitted_values_snaive,.fitted,colour="blue",linetype="dashed")
#mean model
fitted_values_mean <- fit_mean %>% augment()
head(fitted_values_mean)
forecast_mean <- fit_mean %>% forecast(h=7)
forecast_mean %>% autoplot(scs_pedestrians,level=FALSE)+
autolayer(fitted_values_mean,.fitted,colour="blue",linetype="dashed")
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
library(fpp3)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
data<- pedestrian %>% filter(Sensor %in% c("Southern Cross Station", "Bourke Street Mall (North)")) %>% select(Sensor, Date, Count)
data
data <- data %>% group_by(Sensor) %>% index_by(day= Date) %>% summarise(mean_count = mean(Count))
data=data %>% filter(day>"2016-07-01") %>%filter(day>"2016-10-25")
data
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
autoplot(scs_pedestrians, Count) +
scale_x_date(date_breaks = "5 weeks",
minor_breaks = "1 week") +
theme(axis.title.x = element_text(angle=90))
scs_pedestrians$mean_val = mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians <- scs_pedestrians %>% mutate(`trend_class` = slider::slide_dbl
(Count,mean, .before = 3, .after= 3, .complete = TRUE)
)
scs_pedestrians
scs_pedestrians %>%  autoplot(Count) +
geom_line(aes(y= `trend_class`), linetype = "dashed") +
geom_line(aes(y= `mean_val`), linetype = "dashed")
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians
fit <- scs_pedestrians %>% model(mean = MEAN(Count))
fit
fit2 <-scs_pedestrians %>% model(Drift = RW(Count ~ drift()))
fit2
fit3<- scs_pedestrians %>% select(Count) %>% model(snaive = SNAIVE(Count))
fit3
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
forecasts1 <- fit3 %>% forecast(h=7)
forecasts1
forecasts2 <- fit %>% forecast(h=7)
forecasts2
autoplot(forecasts1)
autoplot(forecasts2)
library(fpp3)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
q1 <- pedestrian %>% filter(Sensor == "Southern Cross Station" | Sensor == "Bourke Street Mall (North)",
Date >= "2016-07-01" & Date <= "2016-10-25")
q1
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
q1 %>%
autoplot(Count) +
scale_x_datetime(date_breaks = "5 weeks", minor_breaks ="1 week")
# There is not another sensor in the database
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
q1 %>% filter(Sensor == "Southern Cross Station") %>% pull(Count) %>% max()
q1 %>% filter(Sensor== "Bourke Street Mall (North)") %>% pull(Count) %>% max()
q2 = q1 %>% filter(Sensor== "Southern Cross Station", Count == 3474) # date is 2016-08-02
q3 = q1 %>% filter(Sensor== "Bourke Street Mall (North)", Count == 4812) # date is 2016-08-31
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians %>%
autoplot() + scale_x_date(date_breaks ="5 weeks", minor_breaks = "1 week")
scs_pedestrians$mean_val = mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians <- scs_pedestrians %>%
mutate(
trend_class = slider::slide_dbl(Count, mean, .before = 3, .after = 3, .complete = TRUE)
)
scs_pedestrians %>%
autoplot(Count, colour = "gray") +
geom_line(aes(y = mean_val), colour = "blue", linetype = "dashed") +
geom_line(aes(y = trend_class), colour = "darkgreen", linetype = "dashed")
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
# all models
fit <-
scs_pedestrians %>%
model(
mean = MEAN(Count),
SNaive = SNAIVE(Count),
Drift = RW(Count ~ drift())
)
fitted_vals <- fit %>% augment()
fitted_vals
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
forecasts <- fit %>% forecast(h=7) # h = 7 since our data goes by days, and our seasonal period is one week
forecasts
fitted_valsSNaive <- fitted_vals %>% filter(.model == "SNaive")
fitted_valsDrift <- fitted_vals %>% filter(.model == "Drift")
forecasts %>%
filter(.model == "SNaive") %>%
# autoplot()
autoplot(scs_pedestrians, colour = "gray") +
autolayer(fitted_vals, .fitted, colour = "blue", linetype = "dashed")
forecasts %>%
filter(.model == "Drift") %>%
# autoplot()
autoplot(scs_pedestrians, colour = "gray") +
autolayer(fitted_vals, .fitted, colour = "blue", linetype = "dashed")
library(fpp3)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
pedestrian_adjust <-
pedestrian %>%
filter(Sensor %in% c("Southern Cross Station", "Bourke Street Mall (North)")) %>%
filter(Date >= "2016-07-01", Date <= "2016-10-25")
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
pedestrian_adjust %>%
#as_tsibble(index = Date, key = c(Sensor)) %>%
autoplot(Count) #+ scale_x_date(breaks = "5 week", minor_breaks = "1 week")
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
BSM <-
pedestrian_adjust %>%
filter(Sensor == "Bourke Street Mall (North)") %>%
arrange(desc(Count)) %>%
pull(Date)
BSM_date <- BSM[1]
SCS <-
pedestrian_adjust %>%
filter(Sensor == "Southern Cross Station") %>%
arrange(desc(Count)) %>%
pull(Date)
SCS_date <- SCS[1]
BSM_date
SCS_date
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
library(patchwork) # Used to manage the relative location of ggplots
library(GGally)
scs_pedestrians %>%
autoplot() + scale_x_date(breaks = "5 week", minor_breaks = "1 week")
scs_pedestrians$mean_val = mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians_trend <-
scs_pedestrians %>%
mutate(
`trend_class` = slider::slide_dbl(Count, mean,
.before = 3, .after = 3, .complete = TRUE)
)
scs_pedestrians_trend %>%
autoplot(`trend_class`) +
geom_line(aes(y = mean_val),linetype = "dashed")
scs_pedestrians_trend
scs_pedestrians_trend %>%
autoplot(Count, colour = "gray") +
scale_x_date(
minor_breaks = "1 week",
breaks = "5 weeks"
) +
geom_line(aes(y = trend_class), colour = "#D55E00") +
geom_line(aes(y = mean_val), colour = "red", linetype="dashed") +
# Flip x-labels by 90 degrees
theme(axis.text.x = element_text(angle = 90))
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
fit <-
scs_pedestrians %>%
model(
mean = MEAN(Count),
drift = RW(Count ~ drift()),
snaive = SNAIVE(Count)
)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
forecast <-fit %>% forecast(h=7)
forecast %>%
filter(.model == "snaive") %>%
autoplot(scs_pedestrians)
forecast %>%
filter(.model == "mean") %>%
autoplot(scs_pedestrians)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians %>%
ACF(Count) %>%
autoplot()
fit %>%
select(mean) %>%
gg_tsresiduals()
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
scs_pedestrians %>%
ACF(Count) %>%
autoplot()
fit %>%
select(mean) %>%
gg_tsresiduals()
library(fpp3)
library(ggplot2)
library(dplyr)
library(tsibble)
library(forecast)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
filtered_data <- pedestrian %>%
filter(Sensor %in% c("Southern Cross Station", "Bourke Street Mall (North)") &
Date >= as.Date("2016-07-01") & Date <= as.Date("2016-10-25")) %>%
group_by(Sensor, Date) %>%
summarise(Total_Pedestrians = sum(Count))
print(filtered_data)
library(fpp3)
library(ggplot2)
library(dplyr)
library(tsibble)
library(forecast)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
filtered_data <- pedestrian %>%
filter(Sensor %in% c("Southern Cross Station", "Bourke Street Mall (North)") &
Date >= as.Date("2016-07-01") & Date <= as.Date("2016-10-25")) %>%
group_by(Sensor, Date) %>%
summarise(Total_Pedestrians = sum(Count))
print(filtered_data)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
filtered_data <- pedestrian %>%
filter(Sensor %in% c("Southern Cross Station", "Bourke Street Mall (North)") &
Date >= as.Date("2016-07-01") & Date <= as.Date("2016-10-25")) %>%
group_by(Sensor, Date) %>%
summarise(Total_Pedestrians = sum(Count))
max_pedestrians_per_sensor <- filtered_data %>%
group_by(Sensor) %>%
top_n(1, Total_Pedestrians)
print(max_pedestrians_per_sensor)
filtered_data <- pedestrian %>%
filter(Sensor %in% c("Southern Cross Station", "Bourke Street Mall (North)") &
Date >= as.Date("2016-07-01") & Date <= as.Date("2016-10-25")) %>%
group_by(Sensor, Date) %>%
summarise(Total_Pedestrians = sum(Count))
filtered_data
scs_pedestrians <- read.csv("scs_daily_pedestrian.csv") %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index = Date)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
ggplot(scs_pedestrians, aes(x = Date, y = Count)) +
geom_line() +
scale_x_date(
date_breaks = "5 weeks",
minor_breaks = "1 week")
scs_pedestrians$mean_val = mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
mean_val <- mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians <- scs_pedestrians %>%
mutate('5-MA' = as.numeric(slider::slide_dbl(Count, mean, .before = 2, .after = 3, .complete = TRUE)),
'1-MA' = as.numeric(slider::slide_dbl('5-MA', mean, .before = 1 , .after = 0, .complete = TRUE)))
ggplot(data = scs_pedestrians, aes(x = Date)) +
geom_line(aes(y = Count), colour = "gray", size = 0.5) +
geom_line(aes(y = '1-MA'), colour = "blue")
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
mean_val <- mean(scs_pedestrians$Count, na.rm = TRUE)
scs_pedestrians <- scs_pedestrians %>%
mutate('5-MA' = as.numeric(slider::slide_dbl(Count, mean, .before = 2, .after = 3, .complete = TRUE)),
'1-MA' = as.numeric(slider::slide_dbl('5-MA', mean, .before = 1 , .after = 0, .complete = TRUE)))
ggplot(data = scs_pedestrians, aes(x = Date)) +
geom_line(aes(y = Count), colour = "gray", size = 0.5) +
geom_line(aes(y = '1-MA'), colour = "blue")
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
fit_mean <- scs_pedestrians %>% model(mean = MEAN(Count))
fit_drift <- scs_pedestrians %>% model(Drift = RW(Count ~ drift()))
fit_snaive <- scs_pedestrians %>% model(Naive = NAIVE(Count))
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
forecast_mean <- fit_mean %>% forecast(h=1)
forecast_snaive <- fit_snaive %>% forecast(fit_snaive, h = 1)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
forecast_mean <- fit_mean %>% forecast(h=1)
forecast_snaive <- fit_snaive %>% forecast(fit_snaive, h = 1)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
forecast_mean <- fit_mean %>% forecast(h=1)
#
forecast_snaive <- fit_snaive %>% forecast(h = 1)
autoplot(forecast_mean)
autoplot(forecast_snaive)
## YOUR CODE GOES HERE
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
# FEEDBACK: h=1 means one timestep ahead, meaning one day. To produce
# forecasts of up to 1 week ahead you need h=7
forecast_mean <- fit_mean %>% forecast(h=1)
forecast_snaive <- fit_snaive %>% forecast(h = 1)
autoplot(forecast_mean)
autoplot(forecast_snaive)
library(fpp3)
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
pedestrian_data <- read.csv(file.choose("scs_daily_pedestrian.csv"))
library(fpp3)
scs_pedestrians <-
read.csv(file.choose()) %>%
mutate(Date = as.Date(Date)) %>%
as_tsibble(index=Date)
pd<-scs_pedestrians
## YOUR CODE GOES HERE
pd%>%autoplot(Count)+
scale_x_yearweek(date_breaks='5 week',minor_breaks='1 week')
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
pd$mean_val = mean(pd$Count, na.rm = TRUE)
pd
## YOUR CODE GOES HERE
trendclass<-pd%>%
mutate("trend_class"=slider::slide_dbl(mean_val, mean,
.before=3, .after=3, .complete=TRUE)
)%>%
select(Sensor,Date,Count,mean_val,`trend_class` )
trendclass
trendclass%>% autoplot(mean_val)+
geom_line()
## DO NOT CREATE ADDITIONAL CODE SNIPPETS, KEEP EVERYTHING IN A SINGLE SNIPPET.
pd