yrs_predict: Predict license buying
In southwick-associates/lifetime: Provide functions for lifetime pricing

Description Usage Arguments Functions See Also Examples

Convenience functions for predicting on a dataset using a specified model fit.

1
2
3

yrs_predict_retain(predict_df, model_fit, method = "retain")

yrs_predict_renew(predict_df, model_fit, method = "renew")

`predict_df`	data frame to use for prediction
`model_fit`	model to run `predict`
`method`	character name of method used (for plotting)

yrs_predict_retain: Predict license buying - retention
yrs_predict_renew: Predict license buying - renewal

Other functions to estimate annual license buying: yrs_avidity, yrs_calc_avg, yrs_calc, yrs_fit, yrs_lifetime, yrs_plot, yrs_predict_avg, yrs_result, yrs_zero

library(dplyr)
library(ggplot2)
data(all_sports)

df_split <- all_sports %>%
    yrs_avidity() %>%
    yrs_zero_split() %>%
    yrs_zero_filter(function(x) filter(x, age_year %in% 30:50))

# log trend fit of retention rates
train_df <- yrs_calc_retain(df_split)
model_fit <- yrs_fit_retain(train_df)
out <- yrs_predict_retain(data.frame(years_since = 1:40), model_fit)
sum(out$pct) # estimated annual license purchases over 40 years
ggplot(out, aes(years_since, pct)) + geom_line() + geom_point(data = train_df)

# logistic regression based on individual-level renewal
train_df2 <- yrs_calc_renew(df_split)
model_fit2 <- yrs_fit_renew(train_df2)
predict_df2 <- filter(df_split$year0, year == 2011) %>% select(num_years_held)
out2 <- 1:40 %>%
    sapply(function(i) mutate(predict_df2, years_since = i), simplify = FALSE) %>%
    bind_rows() %>%
    yrs_predict_renew(model_fit2)
sum(out2$pct)
ggplot(out2, aes(years_since, pct)) + geom_line() + geom_point(data = train_df)