Different Time Granularities

Last updated on 2025-03-03

Overview

xP&A allows you to have different time granularities in your models and visuals. Granularity refers to the level of detail in which data of a model are organized. Adjusting granularity affects how detailed the data representation is, with finer granularity providing more detailed information and coarser granularity offering a broader overview.

If you have different time granularities in your models and/or visuals, this article provides information on how xP&A determines how the more granular values are 'rolled up' into the higher-level granularity (e.g. weekly to monthly).

Rolling up lower-level granularities to higher-level granularities

Weeks -> months, quarters or years

By default, a week is rolled up to the month, quarter, or year based off the start date of the week.

Example: If I have a week that starts on 30th May, 2022, the full week will count towards the May month, despite 5/7 days of that week being in the month of June.

It is possible to have "custom calendars", i.e. to tell xP&A how you want each week to be rolled up into months. This is useful, for example, if you follow 4/4/5 accounting cycles. At the moment custom calendars are not a user-facing setting, but if you would like to turn one on, please contact us via live-chat in-app.

 

Days -> weeks

By default, days are rolled up into weeks, starting from the start day of the model.

Example: If the first day of the model is 1st July, 2022, then week one is the 7 days beginning 1st July, 2022.

 

Days -> months

By default, daily models are rolled up into months based on the month that the day falls into (as you would expect). So 30th May, 2022 rolls up into May.

  • The default logic xP&A uses is to sum the values from the lower level granularity (e.g. weekly) to roll-up into the higher-level granularity (e.g. monthly), but you can choose other options in the Time Aggregation setting for each variable (e.g. Average, Final..). For more information, see Aggregation Funtions.
  • The "roll-up" logic described below applies to:    
    • Visuals (e.g. how weekly data is rolled up to months or quarters in a table or chart)
    • Models of different granularities (e.g. how a weekly model translates to a monthly model).
Rolling down higher-level granularities to lower-level granularities

Translating higher-level granularities to lower-level granularities is done by division. So to get an annual variable of "120" into a monthly model, 120 is divided 12, to get a monthly value of 120/12 = 10.

Note: This is only relevant for bringing variables between different time-granularity models, this transformation is not possible on charts/tables/other visuals.

Related articles