acf_by_group: Calculate Auto Correlations by Group
In mattmills49/modeler: Exploratory Data Analysis and Modeling Helper Functions

Description Usage Arguments Details Value Examples

acf_by_group takes a data frame with a grouping column and returns a data frame with the autocorrelation values for each group.

1 2	acf_by_group(.data, .group_var, .value_var, ..., .ci = c("none", "white_noise", "ma"), .cl = 0.95)

`.data`	the data frame the data is contained in
`.group_var`	the unquoted variable to be used as the grouping variable
`.value_var`	the name of the time series value to use
`...`	extra parameters to pass to the `acf` function (see `?stats::acf`)
`.ci`	the type of confidence interval to use, defaults to `none`
`.cl`	confidence level for the confidence intervals, defaults to .95

Each row in the resulting data frame will refer to a group-lag auto-correlation value.

white_noise will return confidence intervals for the series assuming the autocorrelations come from a white noise distribution. ma will apply the correction needed if the autocorrelations have some moving average correlation.

a data frame with the acf values by group and lag

pres_ratings <- data.frame(approval = as.numeric(presidents), pre_1965 = c(rep(1, 60), rep(0, 60)))
ratings_acf <- acf_by_group(pres_ratings[!is.na(pres_ratings$approval), ], pre_1965, approval)
## pass params to acf
ratings2_acf <- acf_by_group(pres_ratings, pre_1965, approval, na.action = na.pass)