dhis2-users team mailing list archive

[Jason Pickering <jason.p.pickering@xxxxxxxxx>]

Hi Bob,

Will not get into the semantics here. :-|

Without really knowing any details of the data , beyond what Knut has
shared with us, but it certainly seems more like "Event" data than the sort
of routine data which the current data entry UI is geared to. Our use case
is somewhat more in the middle, meaning that we conduct surveillance for a
certain period of time, and then stop. So, we might have data for Jan-June
for one Orgunit A and Feb-August for Orgunit B. What we would like to see
however when we look at "August" in the Pivots/GIS/etc would be the latest
values, i.e. June data for the first orgunit in the example and August data
for the second orgunit.  It would really not make sense to enter data for
July and August for for Orgunit A first because we have none really and if
we did, it would be repetitive.

Knut in your case, it would seem to make little sense to enter data for
months for which it does not exist, just to be able to visualize it.
However, as is often the case with event type of data, what matters in some
analyses is the data which was last observed. I think this is the
difference between what you are proposing and how we handle it here. What
we need is an aggregation method for the time dimension which retrieves the
latest data only for a given orgunit/datalement/categorycombo.

I certainly see the need for being able to enter data for a single orgunit
for multiple periods though. Some data, like population figures, need to be
entered over many time periods, and from a workflow perspective, would
simplify things a bit.

Regards,
Jason



On Sat, Feb 15, 2014 at 7:13 PM, Bob Jolliffe <bobjolliffe@xxxxxxxxx> wrote:

> Hi Knut
>
> At the risk of appearing even more ignorant, doesn't this scenario bear
> some resemblance to tracked-entity-event, particularly as we have moved to
> generalize this beyond the scope of tracked persons?  Soi event 1, some
> time in September, is a measles "outbreak" in a district.  Event 2, some
> months later, signals the end of the "outbreak" in that district.
>
> Looking at tracker logic and seeing an outbreak-inception-incursion as
> something similar to a stage-in-a-program is perhaps an odd thought but it
> could be that the model works it is just the words which are wrong.
>
> Maybe thats an off the wall thought, but couldn't help sharing ... :-)
>
> Bob
>
>
> On 15 February 2014 07:16, Knut Staring <knutst@xxxxxxxxx> wrote:
>
>> Dear DHIS2 community, looking for good ideas for a use case which is a
>> bit unusual.
>>
>> We'd like to map epidemic outbreaks, and the data are very simple:
>> Outbreak or not, in other words, Yes/No per orgunit (a province or
>> district). Perhaps Yes Only would be best here?
>>
>> The tricky part is the time dimension: We need to produdce maps of
>> affected districts at regular intervals, at least monthly, possibly also
>> more frequent.
>> However, there is a lot of diversity in the data: Sometimes we have a
>> start and end date, sometimes just a number of months.
>>
>> Currently, DHIS2 does not support one screen for multiple periods
>>
>> I see two possibilities:
>>
>> 1) Create just one dataelement "Outbreak" as Yes Only in a monthly
>> dataset. Then one would have to select every affected month and say YES
>>
>> 2) Create two additional data elements: "Start date" and "End date" in a
>> YEARLY dataset. Then also write a script behind the scenes which populates
>> the Outbreak dataelement for the affected months - or even extend it to days
>>
>> Better ideas?
>>
>> --
>> Knut Staring
>> Dept. of Informatics, University of Oslo
>> +4791880522
>> http://dhis2.org
>>
>> _______________________________________________
>> Mailing list: https://launchpad.net/~dhis2-users
>> Post to     : dhis2-users@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~dhis2-users
>> More help   : https://help.launchpad.net/ListHelp
>>
>>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~dhis2-users
> Post to     : dhis2-users@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~dhis2-users
> More help   : https://help.launchpad.net/ListHelp
>
>