This was an idea I have had in my head for months:
I think like...this is not an achievable goal to be honest but a related aspect of this would be that....hmmm the create public outputs pipeline is static from the create visuals in the insight panels code.
This means, and I am not a user so I dont know how bad this is, but this means that if you want to tweak a plot then you need to either edit your data externally or change the create visuals code then rerun vertex
Would it not be sick if you could adjust your settings in app and hit a save outputs button that would then rwite the public outputs that would reproduce that current graph.
a WYSIWG snapshot right where you have one click generation of a snapshot of what you are looking at in analysis mode, obviously changing plot settings is hard but applying the filters and controls to the data isn't?
We could pre-apply filter/control state to df_map (and related dataframes) before calling create_visuals, then all existing panel code runs unchanged but now on a filtered cohort. You could then in like 4 clicks filter it down to a country or down to age between 30 and 40.
but again not sure if that is actually useful from a users point of view or whether what im doing is saving them the 20 seconds it would take to filter the patient level data instead and I don't know its a large amount of work and I often worry that its easy to get enthused about "damn would it not be cool if" rather than on actual useful stuff
This was an idea I have had in my head for months:
I think like...this is not an achievable goal to be honest but a related aspect of this would be that....hmmm the create public outputs pipeline is static from the create visuals in the insight panels code.
This means, and I am not a user so I dont know how bad this is, but this means that if you want to tweak a plot then you need to either edit your data externally or change the create visuals code then rerun vertex
Would it not be sick if you could adjust your settings in app and hit a save outputs button that would then rwite the public outputs that would reproduce that current graph.
a WYSIWG snapshot right where you have one click generation of a snapshot of what you are looking at in analysis mode, obviously changing plot settings is hard but applying the filters and controls to the data isn't?
We could pre-apply filter/control state to df_map (and related dataframes) before calling create_visuals, then all existing panel code runs unchanged but now on a filtered cohort. You could then in like 4 clicks filter it down to a country or down to age between 30 and 40.
but again not sure if that is actually useful from a users point of view or whether what im doing is saving them the 20 seconds it would take to filter the patient level data instead and I don't know its a large amount of work and I often worry that its easy to get enthused about "damn would it not be cool if" rather than on actual useful stuff