CLVTools
CLVTools copied to clipboard
bachmannpatrick
Feature allow for aggregated input to create a clvdata object (cbs)
Currently clvdata() requires individual transaction data as input. In some use cases an option to create a clvdata object directly from aggregated data can be very helpful: the cbs table (x, t.x, T.cal) is created manually by the user.
Note: It will be necessary to also adapt summary() and plot() as the they require transactional data that is not aggregated for some statistics/plots.
Also consider the option the add weekly/monthly/yearly aggregated data:
| Id | Month/year | x | Sales |
|---|---|---|---|
| 123456 | Jan 2019 | 13 | 563.47 |
| 123456 | Feb 2019 | 8 | 368.11 |
Does using period aggregated data require that the LL is changed to use interval-censored data? See Bruce Hardie here about pnbd
Does using period aggregated data require that the LL is changed to use interval-censored data? See Bruce Hardie here about pnbd
Yes. However, there is also the "easy way", where it only affects data preparation. In the end we need a cbs for the LL. The problem is that the end-user would need to provide the date of the first and last transaction.