Timeshift dates to compare events or campaigns

Ask the Community
1

we run a couple events per year, each lasting about 2 weeks. I’m trying to compare the event’s volume over time. The first visualization works.

I create an ‘completed_event_day’ dataset dimension, so I can count volume on day 1, 2, 3, etc for each event like this:

CleanShot 2025-03-27 at 08.34.08@2x
CleanShot 2025-03-27 at 08.34.08@2x1810×1284 174 KB

dimension completed_event_day:

  dimension completed_event_day {
    model: stack_assignment_timing
    label: 'Completed on Event Day'
    type: 'number'
    definition: @aql date_diff('day', events.event_begins, stack_assignment_timing.completed);;
  }

metric percent_completed

  metric percent_completed {
    label: "Percent Completed"
    type: "number"
    hidden: false
    description: ""
    definition: @aql ((stack_assignment_timing.t_stacks * 1.0 )/ (stack_assignment_timing.t_stacks | of_all())) ;;
    format: "#,###0.00%"

Desire

Create a running total and running percent by completed_event_day. I tried the obvious approach, but it seems like running_total needs a date dimension as it’s running dimension

CleanShot 2025-03-27 at 08.53.55@2x
CleanShot 2025-03-27 at 08.53.55@2x1704×904 96.6 KB

I also tried running_total(stack_assignment_timing.t_stacks, stack_assignment_timing.completed) b/c completed it the datetime dimension on the table. I got the saem results in this visualization. Running total isn’t summing over the cmpleted_event_day

Advice?

2

Hi @stonematt,

Yes, by default running_total run over all date dimensions in the explore. To make in run along another type of dimension, you must add that dimension to the params. In your case, it would be something like:

running_total(stack_assignment_timing.t_stacks, stack_assignment_timing.completed_event_day)