Years as a cohort interval

[carrieseltzer]

Thank you for creating the cohorts package! I was looking for something to help me analyze donor retention across years and this package and explanation was far and away the simplest thing I found.

However, I wanted to look at years rather than months, so I created a new function with very minor changes that uses lubridate instead of zoo:

cohort_table_year <- function(df, id_var, date)
{
  dt <- dtplyr::lazy_dt(df)
  dt %>% dplyr::rename(id = {
    {
      id_var
    }
  }) %>% dplyr::group_by(id) %>% dplyr::mutate(year = lubridate::year({
    {
      date
    }
  })) %>% dplyr::mutate(cohort = min(year)) %>% dplyr::group_by(cohort,
                                                                 year) %>% dplyr::summarise(users = dplyr::n_distinct(id)) %>%
    tidyr::pivot_wider(names_from = year, values_from = users) %>%
    dplyr::ungroup() %>% dplyr::mutate(cohort = 1:dplyr::n_distinct(cohort)) %>%
    tibble::as_tibble()
}

Feel free to add this (and/or rewrite with zoo instead).