AlishaRuqshan
commented
4 months ago # HW2
# All packages
library(fredr)
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
library(dygraphs)
library(xts)
#> Loading required package: zoo
#>
#> Attaching package: 'zoo'
#> The following objects are masked from 'package:base':
#>
#> as.Date, as.Date.numeric
#>
#> ######################### Warning from 'xts' package ##########################
#> # #
#> # The dplyr lag() function breaks how base R's lag() function is supposed to #
#> # work, which breaks lag(my_xts). Calls to lag(my_xts) that you type or #
#> # source() into this session won't work correctly. #
#> # #
#> # Use stats::lag() to make sure you're not using dplyr::lag(), or you can add #
#> # conflictRules('dplyr', exclude = 'lag') to your .Rprofile to stop #
#> # dplyr from breaking base R's lag() function. #
#> # #
#> # Code in packages is not affected. It's protected by R's namespace mechanism #
#> # Set `options(xts.warn_dplyr_breaks_lag = FALSE)` to suppress this warning. #
#> # #
#> ###############################################################################
#>
#> Attaching package: 'xts'
#> The following objects are masked from 'package:dplyr':
#>
#> first, last
library(forecast)
#> Registered S3 method overwritten by 'quantmod':
#> method from
#> as.zoo.data.frame zoo
library(reprex)
fredr_set_key("052142bc981666b4ebcb1c8df98d006b")
# Reading Data
icnsa_data = fredr(series_id = "ICNSA")
# Date column conversion to date data type
icnsa_data$date <- as.Date(icnsa_data$date)
# Updataing dataframe to remove other columns and keep only date and value
updated_icnsa_df <- icnsa_data %>%
select(date, value) %>%
na.omit()
# Define the COVID-19 period
covid_start <- as.Date("2020-01-01")
covid_end <- as.Date("2022-12-31")
icnsa_no_covid <- subset(updated_icnsa_df, date < covid_start | date > covid_end)
# Autocorrelation Plot for dataframe
acf(icnsa_no_covid$value)
# Partial Autocorrelation Plot for dataframe
pacf(icnsa_no_covid$value)
str(icnsa_no_covid)
#> tibble [2,822 × 2] (S3: tbl_df/tbl/data.frame)
#> $ date : Date[1:2822], format: "1967-01-07" "1967-01-14" ...
#> $ value: num [1:2822] 346000 334000 277000 252000 274000 276000 247000 248000 326000 240000 ...
# Ensuring value column as a numeric datatype
icnsa_no_covid$value <- as.numeric(icnsa_no_covid$value)
# Creating a timeseries object with frequency 7 as weekly data
ts_updated_icnsa_df_no_covid <- ts(icnsa_no_covid$value, frequency = 7)
# Fitting an ARIMA model using auto.arima function. We can also specify p,q,d values using above acf and pcf plots.
arima_model_no_covid <- auto.arima(ts_updated_icnsa_df_no_covid)
summary(arima_model_no_covid)
#> Series: ts_updated_icnsa_df_no_covid
#> ARIMA(2,1,2)
#>
#> Coefficients:
#> ar1 ar2 ma1 ma2
#> 0.9555 -0.0637 -1.1942 0.2068
#> s.e. 0.1222 0.1030 0.1204 0.1179
#>
#> sigma^2 = 2.477e+09: log likelihood = -34510.67
#> AIC=69031.34 AICc=69031.36 BIC=69061.07
#>
#> Training set error measures:
#> ME RMSE MAE MPE MAPE MASE
#> Training set -146.1736 49722.01 33579.3 -1.452661 9.090663 0.4276818
#> ACF1
#> Training set 0.0005209598
# Forcasting the value
forecast_values <- forecast(arima_model_no_covid, h = 1)
plot(forecast_values, main = "ARIMA Forecast",
xlab = "Date", ylab = "Number of Cases")
forecast_mean <- mean(forecast_values$mean, na.rm = TRUE)
print(forecast_mean)
#> [1] 235751.1Created on 2024-02-09 with reprex v2.1.0
- Forecasted value for current week after removing covid data is 235751.1
Created on 2024-02-09 with reprex v2.1.0