RTL (version 1.3.7)

chart_zscore: Z-Score applied to seasonal data divergence

Description

Supports analytics and display of seasonal data. Z-Score is computed on residuals conditional on their seasonal period. Beware that most seasonal charts in industry e.g. (NG Storage) is not de-trended so results once you apply an STL decomposition will vary from the unajusted seasonal plot.

Usage

chart_zscore(
  df = df,
  title = "NG Storage Z Score",
  per = "yearweek",
  output = "zscore",
  chart = "seasons"
)

Value

Time series of STL decomposition residuals Z-Scores, or standard seasonal chart with feast package.

Arguments

df

Long data frame with columns series, date and value. tibble

title

Default is a blank space returning the unique value in df$series. character

per

Frequency of seasonality "yearweek" (DEFAULT). "yearmonth", "yearquarter" character

output

"stl" for STL decomposition chart, "stats" for STL fitted statistics. "res" for STL fitted data. "zscore" for residuals Z-score, "seasonal" for standard seasonal chart.

chart

"seasons" for feasts::gg_season() (DEFAULT) "series" for feasts::gg_subseries()

Author

Philippe Cote

Examples

if (FALSE) {
df <- eiaStocks %>% dplyr::filter(series == "NGLower48")
title <- "NGLower48"
chart_zscore(df = df, title = " ", per = "yearweek", output = "stl", chart = "seasons")
chart_zscore(df = df, title = " ", per = "yearweek", output = "stats", chart = "seasons")
chart_zscore(df = df, title = " ", per = "yearweek", output = "res", chart = "seasons")
chart_zscore(df = df, title = " ", per = "yearweek", output = "zscore", chart = "seasons")
chart_zscore(df = df, title = " ", per = "yearweek", output = "seasonal", chart = "seasons")
}