Open jlivsey opened 5 months ago
First attempt to forecast the weekly initial claims data. I will use the Facebook Prophet.
library(tidyverse)
library(prophet)
library(fredr)
icnsa = fredr(series_id = "ICNSA")
df <- icnsa |>
select(ds = date, y = value)
m <- prophet(df)
#> Disabling weekly seasonality. Run prophet with weekly.seasonality=TRUE to override this.
#> Disabling daily seasonality. Run prophet with daily.seasonality=TRUE to override this.
future <- make_future_dataframe(m, periods = 1)
forecast <- predict(m, future)
tail(forecast[c('ds', 'yhat', 'yhat_lower', 'yhat_upper')])
#> ds yhat yhat_lower yhat_upper
#> 2973 2023-12-23 487058.3 184576.5 791642.2
#> 2974 2023-12-30 544926.0 228351.7 843077.8
#> 2975 2024-01-06 593973.3 294475.9 904214.0
#> 2976 2024-01-13 593983.0 275790.2 904047.3
#> 2977 2024-01-20 539374.0 227671.7 835698.4
#> 2978 2024-01-21 528947.1 232701.8 844609.6
Created on 2024-01-31 with reprex v2.0.2
So I forecast the new data release value will be 528947.1 with confidence interval (232701.8, 844609.6)
Released forecast was 261,029. Error in my point forecast was 528947.1 - 261029 = 267918.1
Released forecast was 261,029. Error in my point forecast was 261,029 - 209106 = 51,923
My select student function
select_student <- function(n = 1){
names_vector <- c("Bedarkar", "Peduru", "Kadiri", "Markunda", "Anand",
"Bhardwaj", "Soni", "Patra", "Maru", "Singh",
"Jaskirat Singh", "Gangu", "Vadhiya", "Kulkarni",
"Samala", "Sarbhai", "Tiwari", "Bhupathiraju",
"Gadhave", "Haveliwala", "Chandrasekar", "Hatote",
"Ghosh", "Prasannan", "Pawar", "Kaveti", "Gudimetla",
"Gunturi", "Yerriboina", "Patil")
sample(names_vector, n, replace = FALSE)
}
A place to document forecasts Google Sheet
Created on 2024-01-30 with reprex v2.0.2
AirPassangers
plot