R Remove multiple seasonality in one time series

Question

I'm dealing with hourly data and my data has 24 hours & 168 hours seasonality.

When I use that code. I removed 24 hours seasonality.

deseasonalized_ec_hourly<-electricity_consumption_ts_hourly/ec_hourly_multip$seasonal

But 168 hours seasonality remains. How can I remove 168 hours seasonality

Answer 1

Here's an example using half-hourly electricity consumption from England and Wals (June - August 2000). The seasonality (daily and weekly) is estimated using the mstl() function from the forecast package. The seasadj() function removes the seasonal components.

library(forecast)

autoplot(taylor)

adjust <- taylor %>% mstl() %>% seasadj()
autoplot(adjust)

^{Created on 2020-05-07 by the reprex package (v0.3.0)}