Forwarding to the dhis-users list, as there is a lot of good info here.

---------- Forwarded message ----------

On Thu, Jun 26, 2008 at 8:18 PM, Knut Staring <[EMAIL PROTECTED]> wrote:

> On Thu, Jun 26, 2008 at 6:30 PM, Paul Amendola <[EMAIL PROTECTED]>
> wrote:
>
>> Hi All
>>
>> Great work, Øystein!! The fix worked perfectly. Greatly appreciated.
>
>
Glad to hear that it worked out and that there was a logical explanation to
this. We are continuously improving the installer and this kind of feedback
is very useful.

>
> Good to hear it's been sorted. Thanks for helping us discover the problem.
>
>>
>> I have some other basic questions, but I was wondering if there's a manual
>> of set of guidelines I can view, in order to save you some time.
>>
>
> Ola is currently updating the user manual - there is A LOT of functionality
> which is sadly currently undocumented. Any help with this (both in regard to
> asking questions and actually documenting how things work) is greatly
> appreciated. It would also be great if you could send them to our common
> developer list: [EMAIL PROTECTED], or even log them in our issue tracker,
> www.hisp.info/dhis2
>

Yes, that's true, I will work on the manual over the next couple of weeks.
For now you can use the Help page which is accessible from the main menu, it
explains the basics, e.g. how to set up indicators.


>
> Either way, here are the questions I'm currently dealing with:
>>
>
>
>
>> - the version I download, is that the typical user interface or the admin
>> model? The older version, 1.4?, had an interface that made data analysis and
>> reports generation really clear. I'm just not sure if there's a similar way
>> to set that up
>
>
Not sure what kind of interface for reporting you are referring to. In the
MS Access based version (1.4) the report component is Excel pivot tables,
which is also there for version 2. You can link the Excel tables to the
dhis2 database using the jdbc:odbc driver. I'll send you an example Excel
file later, where this is set up. Rigth now I am using Ubuntu so I need to
reboot into Windows to set up a pivot table for the Zambia database before I
can send you. Unfortunately we are not satisfied by how OpenOffice Calc
manages pivot tables (data pilots) with larger amounts of data.


>
> The default user you have has complete access to everything (just like the
> online demo), which is certainly not what you want to present to all you end
> users. The User module contains a very fine grained apparatus (also
> underdocumented at the moment) where you can define exactly what each user
> group can see and do, and thus remove access to the administration modules
> for ordinary users.
>
>
In addition to the functionality to filter what the different users can see
and access you can also build new war files with any combination of modules
you need. We use maven for building the war files (web applications) and for
a technical person it is quite easy to modify which modules are part of the
"portal application" by editing a pom.xml file.


>
> - once the db is accessed, is there a way to analyze the data offline?
>
>
> A common way to do this is to hook up Excel pivot tables to the database.
> Ola has created a setup for this and can give you more info.
>

I commented on the pivot tables above. Also note that there it is a quite
common scenario to run DHIS 2 stand-alone in a non-networked environment,
just like you do now with your demo version. The various stand alone
installations can then communicate and synchronise data using xml files and
the import/export module. If you have access to the database, either through
a network or locally on your computer you can use all the data analysis and
reporting functionality that is available through the DHIS 2 user interface.

>
>
>> - is there a level of internet explorer that needs to be used?
>
>
> We have not done very much testing with really old IE versions, but I think
> most of the functionality, except one of the two GIS modules will work fine
> with IE 6. We strongly recommend Firefox or IE 7 though, if at all possible.
>
>
>>
>> - I was wondering if there is a step by step guide to calculating
>> indicators? For example, I'm trying to calculate mortality rate by using the
>> total population and number of deaths
>>
>
> Not sure if we have such a guide - one may exist for the DHIS 1.4 version.
> Perhaps you can also look at the indicators defined in the Zambia base, or
> in the online demo version (which contains demo data from Limpopo in South
> Africa, but with real indicators)
>

If you like you can export both data elements and indicators (metadata
export) using the import/export mdoule, either from your local Zambia
database or the SA sample demo and then you can import what you like (using
import with preview mode) into your empty database. This is a quick way to
get started.

As mentined above, the basic help page explains how indicators are set up.
In the Data Elements and Indicators module you define the indicator formulas
as numerator/denoominator fractions based on the data elements (which you
have to define first).

Indicators are calculated in the datamart module, you can select the
indicators, orgunits and periods you need and all the values will be
calculated and stored in the db table called aggregatedindicatorvalue. One
thing that I am not sure we have covered in the help section yet is
aggregation operator. When indicators are calculated we treat population
data (which often are annual estimates) and routine data (often monthly, e.g
mortality from monthly disease surveillance forms) differently in the
aggregation process. Routine data elements like, Malaria deaths must have
aggregation operator SUM, and population data must have aggregation operator
AVERAGE. So this you have to define at the data element level. Then in the
indicator interface you can define indicator types which are basically the
factor in your formula, 1 for rates, 100 for percentages etc. Indiators can
also be annulaised, meaning that you can calculate an annual estimation
based on one month of routine data (e.g Malaria deaths in September 2007).
If you tick the annualised option the numerator will be multiplied by 12 for
monthky data, 4 for quarterly data etc, but this calculation happens behind
the scene. The best way to learn how this works is to have a look at the
Zambian and the South African (which has extensive indicator
explanations/definitions) databases.

In the data analysis module you will automatically have access to the
indicators that you have exported to the datamart. The Zambian database has
two pivot tables (views9) dfined in the Data Analysis module, and one of
them is called District Indicators. In that one you will have access to all
indicators exported for districts (level 3) in the datamart. Note that there
is only data for August and September 2007. You can go to the Datamart
module and pick the indicators you like, select all orgunits at level 3
(Districts) and then select the periods August and September 2007, and then
export. After that you can open the Data Analysis->District Indicators view
where you'll see that your nee indicators have been added.

Please feel free to contact us for further questions or issues, it is
actually very useful for us to get this kind of feedback from "outsiders".
Typically some of us are also involved in the implementation as well setting
up the databases, so some new ideas and thinking around these concepts are
very welcome. As Knut said  our mailing list (
http://www.hisp.info/mailman/listinfo/dhis-dev) is an excellent place for
these kind of communciation.

Ola
_______________________________________________
Dhis-users mailing list
[email protected]
http://www.hisp.info/mailman/listinfo/dhis-users

Reply via email to