Skip to contents

Cohort Analysis

cohort.Rd

Database-friendly cohort analysis function. A remake of https://github.com/PeerChristensen/cohorts, combining cohort_table_month, cohort_table_year, and cohort_table_day into a single package. Rewritten for database compatibility and tested with Snowflake and DuckDB.

Usage

cohort(
  .data,
  .date,
  .value,
  calendar_type,
  time_unit = "month",
  period_label = FALSE
)

Arguments

.data

tibble or dbi object

.date

date column

.value

id column

calendar_type

clarify the calendar type; 'standard' or '554'

time_unit

do you want summarize the date column to 'day', 'week', 'month','quarter' or 'year'

period_label

do you want period labels or the dates c(TRUE , FALSE)

Value

segment object

Details

  • Groups your .value column by shared time attributes from the .date column.

  • Assigns each member to a cohort based on their first entry in .date.

  • Aggregates the cohort by the time_unit argument (day, week, month, quarter, or year).

  • Computes the distinct count of each cohort member over time.