Dynamic X axis on a boxplot depending on user selection
I am using Qlik Sense (February 2023 Patch 10).
I want to enhance my current boxplot of continuous_var over time_dim, disregarding the settings of filter_var. Currently, I plot Sum({$<filter_var=>} continuous_var) with [time_dim Year] as X-axis dimension.
As whole years might be a bit coarse, I want to dynamically "make a zoom" on the X-axis and change it to [time_dim YearMonth] depending on the selection of a time span by the user (VizLib's Calendar filter, in my case). Using a drilldown dimension won't work if I want to display months spanning two different years. More generally, I want to display data by week for the last 26 weeks, or by month for the last 24 months, or by quarter for the last 24 quarters, or by year.
I already got this to work for a bar chart of Count(ID) over time_dim, but this does not currently work for boxplots. For the bar chart, I use the following expression for the X axis: =$(=if(count(distinct {$<filter_var=>} [time_dim YearMonth]) <= 24, '[time_dim YearMonth]', '[time_dim Year]')) This selects [time_dim YearMonth] instead of [time_dim Year] whenever 24 months or less have been selected.
To avoid empty bars, I use the following expression to replace Count([ID]): =sum({<filter_var=, [time_dim Date]={">=$(=Min({<filter_var=>} [time_dim Date]))<=$(=Max({<filter_var=>} [time_dim Date]))"} > } Aggr(Count([ID]), [time_dim Date])) Here, I aggregated over single dates, since I may want to have boxplots by week as well.
For the boxplot, I imagine that I need to replace continuous_var by something along this line: =sum({...} Aggr(Sum(continuous_var, ID)) But that doesn't do the trick. Any thoughts?
it seems like you've already put considerable effort into customizing your visualizations, here is an attempt towards your goal, but I would like to point out that I didn't test it:
