dhis2-devs team mailing list archive

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

Hi Calle, it may not be good, but it is usually always necessary with the
type of data which DHIS 2 normally stores. Take logistics data as an
example. The data is usually reported each day, but for which one may not
have recorded zeros for particular activities, such as when a given
commodity is not dispensed. It is not really efficient to record all these
zeros, so normally, they would simply not be recorded. When it comes to
calculating a daily stock level, one will have missing observations,
combined with a requirement to have a daily stock level. So, in this case,
you must use the "last observation carried forward" approach, unless you
want to record a zero for each day, for potentially thousands of
commodities. Not really efficient.

Same could be said when aggregating from facilities to national level. If
any data is missing. It would needed to be imputed, either by using the
LOCF approach, or other statistical methods (means for instance). We just
need to take this into account when dealing with cumulative time data,
otherwise, any sorts of calculations are usually simply not possible.

Regards,
Jason




On Thu, Jan 22, 2015 at 6:55 AM, Calle Hedberg <calle.hedberg@xxxxxxxxx>
wrote:

> Hi,
>
> I'm not sure if auto-replicating the previous month's value for a
> missing month is such a good idea. It's conceptually very similar to
> the ability in DHIS 1.x to use regression analysis to insert missing
> values, BUT we always blocked that for recent months - the idea was
> always that using regression analysis to insert missing values was a
> last resort which only could/should be available after several months
> (i.e. after all other avenues to get the missing data are exhausted).
>
> regards
> Calle
>
> On 21/01/2015, Jason Pickering <jason.p.pickering@xxxxxxxxx> wrote:
> > One potentially tricky thing to keep in mind here is the problem of
> missing
> > data, i.e. if the facility does not report for a month, or they report a
> > zero and the data element is zero insignificant. Normally in time-series
> > data, the "last observation would be carried forward" (LOCF). The spec
> does
> > not really mention this, but in all likelihood, this would need to be
> part
> > of the implementation.
> >
> > Regards,
> > Jason
> >
> >
> > On Wed, Jan 21, 2015 at 10:32 AM, John Lewis <johnlewis.hisp@xxxxxxxxx>
> > wrote:
> >
> >> Hi Lars,
> >> Example in Excel. We had this chart in Dhis1 series.
> >> John
> >>
> >> On Wed, Jan 21, 2015 at 4:18 PM, John Lewis <johnlewis.hisp@xxxxxxxxx>
> >> wrote:
> >>
> >>> Hi Lars,
> >>> Done.
> >>>
> >>>
> https://blueprints.launchpad.net/dhis2/+spec/visualisation-cumulative-chart
> >>> Regards,
> >>> John
> >>>
> >>> On Mon, Jan 19, 2015 at 5:11 PM, Lars Helge Øverland
> >>> <larshelge@xxxxxxxxx
> >>> > wrote:
> >>>
> >>>> Hi John,
> >>>>
> >>>> yes that would be possible to implement. Feel free to write a
> blueprint
> >>>> with a bit more detail and we can schedule it on the road map.
> >>>>
> >>>> regards,
> >>>>
> >>>> Lars
> >>>>
> >>>>
> >>>> On Fri, Jan 16, 2015 at 3:00 AM, John Lewis <johnlewis.hisp@xxxxxxxxx
> >
> >>>> wrote:
> >>>>
> >>>>> Hi Guys,
> >>>>> Various countries want to analyse the data cumulatively. for example
> >>>>> compare BCG dose given (data element) for last 12 (or /6/3) month
> with
> >>>>> graph showing monthly data and cumulative value.
> >>>>> Is this possible to include it in analysis
> >>>>> Regards,
> >>>>> John
> >>>>>
> >>>>>
> >>>>> _______________________________________________
> >>>>> Mailing list: https://launchpad.net/~dhis2-devs
> >>>>> Post to     : dhis2-devs@xxxxxxxxxxxxxxxxxxx
> >>>>> Unsubscribe : https://launchpad.net/~dhis2-devs
> >>>>> More help   : https://help.launchpad.net/ListHelp
> >>>>>
> >>>>>
> >>>>
> >>>
> >>
> >> _______________________________________________
> >> Mailing list: https://launchpad.net/~dhis2-devs
> >> Post to     : dhis2-devs@xxxxxxxxxxxxxxxxxxx
> >> Unsubscribe : https://launchpad.net/~dhis2-devs
> >> More help   : https://help.launchpad.net/ListHelp
> >>
> >>
> >
> >
> > --
> > Jason P. Pickering
> > email: jason.p.pickering@xxxxxxxxx
> > tel:+46764147049
> >
>
>
> --
>
>
> *******************************************
>
> Calle Hedberg
>
> 46D Alma Road, 7700 Rosebank, SOUTH AFRICA
>
> Tel/fax (home): +27-21-685-6472
>
> Cell: +27-82-853-5352
>
> Iridium SatPhone: +8816-315-19274
>
> Email: calle.hedberg@xxxxxxxxx
>
> Skype: calle_hedberg
>
> *******************************************
>



-- 
Jason P. Pickering
email: jason.p.pickering@xxxxxxxxx
tel:+46764147049