Yes it initially does look exactly how I want but if you do any sort
of filtering on any of the facets the whole thing disappears.

Initially I only had the 34 records which each had a status 'Active',
'Inactive', 'Slothful' etc but then I added 3 extra records to
represent those states themselves and put in the metadata that status
was an 'item' and not just a string. Note in the first chart that
these 3 records are not plotted but in the middle chart it is only
these three records which are plotted. I want to be able to filter say
on 'Category' and have the subtotal for each status shown that matchs
that category. At the moment if you filter on Category it will filter
out the 3 status records which is the base for the second bar chart
and not the actual real data I want to plot.

If I had the entire set of data (ie one record for every person) then
everything would work fine but I have 120,000+ records which won't
scale which is why I have summarised each group of people into one
record and added the membership count as data.

I guess the bar chart is assuming that each record becomes one bar, so
what I want is to group the records into sets before the bar chart
widget seem it. Like a 'group by' clause in sql but no ideal how to
express this in the exhibit expresion language or even if it has the
expresive power to represent that.

So this would work perfect if I have the full data set:

<div ex:role="view" ex:viewClass="BarChartView" ex:y=".Status"
ex:x="count(.Status)" />

but because I have each record grouped with effectively a weighting I
want something like

<div ex:role="view" ex:viewClass="BarChartView" ex:y="unique(.Status)"
ex:x="add(.Count)" />

Any thoughts? If the expression language doesn't support this how hard
would it be to add? Is there a simpler way to do this that avoids the
whole problem?

Another thought I had was that in my second chart example when you
filter it would work fine if those 3 records were included in the
filtered result set, is there anyway to force them to stay or to not
filter them in the first place?


On Aug 29, 10:16 am, mleden <[email protected]> wrote:
> Hi Brendan,
>
> I guess I must be not quite following you.  I looked at your URL and
> the second view ("Chart with right bars but wrong data") appears to me
> to be representing what you want.  It is subtotaling the values in the
> Count property of your dataset at each Status value (Active, Inactive,
> Slothfull).  Can you describe exactly how you would like the end
> result to look, given the existing dataset?
>
> -Mark
>
> On Aug 27, 9:53 pm, Brendan Heywood <[email protected]> wrote:
>
> > Hi,
>
> > I have a simple data set with three dimensions, Status, Category and
> > Type and each record has a count of the number of people.
>
> > Example:http://www-dev.une.edu.au/bheywood/sandbox/urs-test/
>
> > What I want to do is given a particular dimension / facet (say Status
> > in the example above) is produce a bar chart / pie chart which shows
> > the totals of the counts in each record.
>
> > Ie at the moment there are 34 records and all of them have a Status of
> > either Active or Inactive and I want to merge all of those in the bar
> > chart to show just two bars and the totals of all records with that
> > Status.
>
> > Is this even possible just using expressions in the x / y attributes
> > of the bar chart? I’d still like to be able to then drill down into
> > other facets and get say the chart of Active/Inactive for just a
> > particular category or type of people.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"SIMILE Widgets" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/simile-widgets?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to