Real-world example: Summary in a Tab


(Mark Anderson) #1

This is a simple solution, but I’ve not seen it illustrated. If screen space is short and you want a running total, one solution is to make a tab (view type immaterial) and hoist a note whose display expression shows the stat you want. In my case here it’s the completion state of a loop through some 4.5k Wikipedia threads I’m researching:

Note the second, non-selected tab (top right). This is live current real-world use - for those asking how things might (not should) look. I note the last in part to push home the point that tidiness isn’t as important as getting to grips with the data.

Here it was apparent early on that the sheer size of the task meant I needed several passes. Currently, I’m gathering start/end dates of threads (the Wikipedia API’s too coarse for this sort of task: e.g. ‘last’ isn’t necessarily an edit made months later that isn’t really on topic.). Column view data helps check progress, see missing data and help surface patterns in the data. The many empty user KAs in the text pane await the next pass. If space was pressing, I’d edit the prototype’s [sic] KAs to optimise the set for this pass and then alter before the next pass through the data.

The code for the completion figure is:

$MyNumber2=(find($Prototype=="pRequest"&$IsAlias==false&$UserCount>0).unique.count);
$MyNumber3=(find($Prototype=="pRequest"&$IsAlias==false).count);
$MyNumber=($MyNumber2/$MyNumber3)*100;
$MyString=$Name+" "+$MyNumber.precision(2)+"%";

…Which I store as an edict, as I don’t need these big finds running all the time - updating agents causes an edict refresh.

Why use $MyString? The result of the edict code is to create a literal string that can be called as the Display Expression:

$MyString

No calculation required in generating the string used for the DE.

Now, back to the research…


(Paul Walters) #2

Interesting and very cool.


(Mark Anderson) #3

I meant to update. Here is how i used the above in practice, with the first pass complete:

tab-totals1

The tab tells me I’ve now analysed every one (100%) of 4,278 discrete threads started by 2,044 discrete Wikipedia editors and involving a total of 2,722 discrete editors. Numerous other counts and observations taken along the way can now be exported as a data table for use in R.