dhis2-users team mailing list archive

[Ola Hodne Titlestad <olatitle@xxxxxxxxx>]

Hi again Roger,

I think I have found the source of this lack of roll-up/aggregation problem.
I looked at your database and found that for the data element "Malaria
cases" you had specified Aggregation Levels Region and Country.

Aggregation levels is a complicated feature explained in the user manual,
section 9.2.1.5, but briefly your settings here mean that when asking the
system for an aggregated values for Regions it will not look for any values
below the Region level, and for Country it will not look below the country
level. That is why your roll up does not work. It seems you have
misunderstood the use of Aggregation levels.

What you need to do is to edit the data element and remove the selected
aggregation levels sio that you leave that selection box blank. We try to
hide this feature, and these levels only show when you tick the Aggregation
level box. Then the system will follow the default behaviour for aggregation
which is to include all data at all levels when aggregating up the org
hierarchy, in your case it will then find the malaria case values you have
registered on the district level and sum up these to regional totals.

The aggregation levels setting in data elements do not have anything to do
with which levels that can produce aggregate data, that you specify in the
various report tools (like report tables), but in stead specifies where to
look for data when aggregating up the org hierarchy.

This section in the UI definitely needs a help icon and we could probably
try to find better names and more explanatory text for this feature.  Sorry
for the confusion.

----------------------------------
Ola Hodne Titlestad (Mr)
HISP
Department of Informatics
University of Oslo

Mobile: +47 48069736
Home address: Vetlandsvn. 95B, 0685 Oslo, Norway. Googlemaps
link<http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=Vetlandsvn.+95B,+0685+Oslo,+Norway>



On 26 August 2010 11:47, Ola Hodne Titlestad <olatitle@xxxxxxxxx> wrote:

> On 26 August 2010 00:28, Friedman, Roger (CDC/CGH/DGHA) (CTR) <
> rdf4@xxxxxxx> wrote:
>
>>   Folks --
>>             While I am here, I read Ola's chapter on reporting, which was
>> very useful.  It made clear what I was just coming to realize, that
>> disaggregation groups are not rolled up to the next level.  This means that
>> indicators based on disaggregation groups (percent of patients whose CD4
>> count was less than 350 at the first visit) don't roll up, we are basically
>> left with making each cell of disaggregated data a data element of its own.
>> I suppose I can work around it, but it's a big disappointment.  Am I
>> understanding this right?  Is it on the workplan?
>>
>> Thanks all, Roger
>>
>>
> Hi Roger,
>
> I am not 100% sure I understand what functionality you are asking for, but
> let me give you an example of what is possible and maybe that will help:
>
> I have a data element called "Malaria new cases" and a categorycombo called
> 'MorbidityAge+Gender', two categories MorbidityAge(<5y,>5y) and Gender
> (Male, Female), with the possible catoptioncombos being: <5y Male,<5y
> Female, >5y Male,>5 Female. These data values are collected monthly.
>
> I also have a data element called "Population <5y" with the default (none)
> catcombo collected yearly.
>
> Then I have an indicator called 'Malaria incidence rate <5y' with a
> numerator expression like:
> ("Malaria new cases"+"<5y Male") + ("Malaria new cases"+"<5 Female")
>
> and a denominator expression: "Population <5y"
>
> The indicator is of type percentage and annualised (makes the factor 1200
> for monthly indicator values).
>
> In any of the report tools (report table, chart, map, datamart) I can ask
> for the indicator value of "Malaria incidence rate <5y" for any orgunit at
> any level from individual health facilities to the country level aggregate
> value.
>
> If this is what you mean by rolling up to the next level then this is
> supported.
>
> Maybe you could give a more detailed example of what you want?
>
> What annoys be a bit and which will be improved for 2.0.6 is that you
> cannot use subtotals, like Malaria new cases <5y directly in the indicator
> expressions, but in stead have to do it piece by piece as in my example
> above. Thsi becomes very tedious if you have many catoptioncombos that needs
> to be added together. As I said, the 2.0.6 will support the use of subtotals
> and totals in such expressions:
> https://blueprints.launchpad.net/dhis2/+spec/subtotals-in-expressions
>
>
> ----------------------------------
> Ola Hodne Titlestad (Mr)
> HISP
> Department of Informatics
> University of Oslo
>
> Mobile: +47 48069736
> Home address: Vetlandsvn. 95B, 0685 Oslo, Norway. Googlemaps link<http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=Vetlandsvn.+95B,+0685+Oslo,+Norway>
>
>