est_churn: Estimate churn by year from license history

In southwick-associates/salic: Prepare Agency Dashboard Data

Description

This function requires a correctly formated history table (see history). It runs a mean of the lapse value (per year), optionally by segment (and also shifts year forward by 1 so that churn in current year reflects lapse pct from last year). It also runs a validation test: pct change per year.

Usage

est_churn(history, segment = "tot", test_threshold = 30,
  show_test_stat = FALSE, suppress_warning = FALSE, outvar = "churn")

Arguments

history	data frame: input license history table
segment	character: defaults to "tot", which indicates no segmentation. Alternatively specifiy other license history variables ("res", "sex", etc.)
test_threshold	numeric: threshold in whole number percentage points for pct change per year. A warning will be printed if the absolute value of the change for any year exceeds the threshold.
show_test_stat	logical: If TRUE, the output table will include a variable holding the test statistic for each row.
suppress_warning	logical: If TRUE, no test warning will be displayed (even if threshold is exceeded). Test statistics can still be included by setting show_test_stat = TRUE.
outvar	character: name of variable that stores metric

See Also

Salic Function Reference: salic

Other dashboard functions: check_threshold, est_part, format_result, scaleup_part

Examples

library(dplyr)
data(history)
history <- history %>%
    label_categories() %>%
    recode_agecat() %>%
    filter(!agecat %in% c("0-17", "65+"))
est_churn(history)

# apply across all segments
segs <- c("tot", "res", "sex", "agecat")
sapply(segs, function(x) est_churn(history, x), simplify = FALSE)

southwick-associates/salic documentation built on Nov. 5, 2019, 9:13 a.m.