dhis2-devs team mailing list archive

[Bob Jolliffe <bobjolliffe@xxxxxxxxx>]

On 23 May 2011 12:19, Juma Lungo <juma.lungo@xxxxxxxxxxxx> wrote:

>  Let me try to define the problem again. MTUHA is the SWAHILI name of the
> HMIS, thus MTUHA Report is HMIS report. We want to generate HMIS reports by
> using the DHIS2. However, we want DHIS2 to receive aggregated data from
> Care2x database (instead of those values be entered straight to the DHIS2).
> One aggregate data value of a DHIS2 can come from one or more Care2x
> database tables and fields.
>
> Thus, we need to identify, which reports we want to generate from DHIS2.
> then from those report, we develop a mapping which indicates where those
> data can be retrieved from the Care2x Database.
>
> I remember this mapping was already done!
>

Right.  This is what I was getting at.  So you know which dataelements you
need to produce the MTUHA report?  And you know how to calculate those from
the data in the care2x database.  If that mapping is already done it is
good.  How does it look?

>
> So now i think we need to go ahead and program ETL tool. The tool should
> compute the Care2x data and insert the values straight to the DHIS database
> as if they were entered by a data clerk.
>

I am not sure if you really want to be reaching into one database and
directly manipulating the other database.  That might be ok for one off
migration or bootstrapping operations, but it is better to import data into
dhis through its web interface than to depend on this sort of
behind-the-scenes subversive operation.   As I suggested before, I think you
should use kettle to generate xml reports from the care2x database and
import these into dhis.

If you have the mappings you speak of above I could assist in defining the
xml reports - I think preferably in a standard SDMX-HD format, but any row
based xml which is using the right codes should work.

Regards
Bob


> Lungo
>
>
>
> On 05/23/2011 02:04 PM, Maurizio Bricola wrote:
>
> Hi all,
>
>  I fully agree with Bob, when he is saying:
>
>  *you can use it (kettle) to interactively fiddle with these things to
> better understand what it is possible to generate off your database.  But It
> should probably be your aim to do this ultimately with php scripts if you
> want to settle the process into a regular workflow.  The php scripts can be
> better integrated into your care2x.  But some programming required :-)
> Perhaps you will end up with a phased approach, starting with kettle and
> gradually folding the queries into a care2x module.*
> *
> *
> The above is the technical advice.
>
>  At this stage there is not much theory to do we should keep attacking
> this issue in a very practical way.
> *
> *
> Basically we need someone with knowledge of the MTUHA report who will put
> some effort in learning Kettle and start to create the report.
>
>  Do we have a RACI Matrix? if not we should think to jointly develop one,
> even from distance using a Google Docs, for example.
>
>  RACI Matrix will help us to have a framework and last but not least the
> job done :)
>
>  RACI means: Responsible, Accountable, Consulted, Informed
>
>  What we need is the following:
>
>  A work breakdown: basically what needs to be done to achieve the goal,
> and should be specific enough to answer the question: "Who does X?"
>
>  A list of roles like: Project Manager, Programmers, etc
>
>  Another tool we could use is the Who/What/When Matrix in this case it is
> important to start with the Who, to answer the question: "Who does what
> when?"
>
>  Either one or the other should be jointly developed in
> a participatory way since *we all share the responsibility* to reach our
> objective.
>
> In case we already have something like that, it might me good to share it,
> review it and refresh it. Colleagues which did not have the time to fully
> commit because of different reasons they might have the opportunity to give
> their valuable contribution now.
>
>  With Kind Regards,
>
> Maurizio Bricola
> Technical Advisor (Kenya - Malawi - Tanzania)
> International Institute for Communication and Development (IICD)
> P.O. Box 11586, 2502 AN The Hague, the Netherlands
> Visitors: Raamweg 5, 2596 HL The Hague
> Phone: +31-(0)70-311 73 11 Mobile: +31-(0)6 33 77 35 41
> Website: www.iicd.org
>
>
>
>
>
>
>
>
> On 23 May 2011 12:36, Mauri Niemi <mauri.niemi@xxxxxxxxx> wrote:
>
>> Hi All,
>>
>>  When i first time saw DHIS and Vincent visited our team in Arusha, I
>> think it was 2006 I liked the flexibility of DHIS and since then I wanted
>> that we could use it as general reporting tool for Care2x. It should cover
>> mandatory reports for government (Mtuha) and also hospital specific reports.
>> The idea was that we would not need programmers but people who know the
>> database and can make queriees and produce reports which are then  analyzed
>> using DHIS.
>> Care2x does not collect all data needed for Mtuha, initially we thought
>> some parts are easer to do manually (general information about health
>> facility) and some modules we do not have yet in Care2x (RCH). It is also
>> collecting much data not required by Mtuha (mainly financial data), which
>> most hospitals see very importan. But probaly it is better that webERP
>> acconting packet is reporting financial issues.
>> From this background programmers can decide if php scripts should be used
>> or if it is better to use Kettle or similiar tool.
>>
>>  Best Greetings
>> Mauri Niemi
>>
>>  2011/5/23 Bob Jolliffe <bobjolliffe@xxxxxxxxx>
>>
>>> Hi Lungo (cc'd to list)
>>>
>>> Can you make this discussion more concrete by by providing details of the
>>> MTUHA report?
>>>
>>> Taking things one step at a time, can you then also verify that dhis2 in
>>> TZ has all the existing required dataelements for this report.  Without more
>>> detail I don't know if the MTUHA report is a report of dataelements or a
>>> processed report of indicators.  If the latter then you should identify a
>>> report of dataelements required to calculate the MTUHA indicators.
>>>
>>> I would suggest that we then define an SDMX-HD DSD for this report (ie
>>> specify all the codes to be used for dataelements, dimensions etc) and
>>> ensure that DHIS can consume datasets conforming to this.  If all the
>>> required dataelements exist within the TZ dhis database then I have a small
>>> script which can assist in generating this from the dhis metadata.
>>> Afterwards it can be trimmed and tidied a bit.
>>>
>>> Then there is the small(!) matter of producing this dataset from the data
>>> within the care2x database.  Most fundamentally this will be about
>>> performing SQL queries on that database and generating XML from the
>>> resultsets.  I don't think it matters too much whether those queries are
>>> executed within the context of kettle or a php script.  What is more
>>> important is to understand whether the the required queries are possible
>>> from the existing data.  It might be you are only able to produce a subset
>>> of what you need for example.  You would be in a better position to know
>>> this.  As I say, without better knowledge of the MTUHA report its difficult
>>> to be too specific.
>>>
>>> I think that if you are familiar with kettle you can use it to
>>> interactively fiddle with these things to better understand what it is
>>> possible to generate off your database.  But It should probably be your aim
>>> to do this ultimately with php scripts if you want to settle the process
>>> into a regular workflow.  The php scripts can be better integrated into your
>>> care2x.  But some programming required :-)  Perhaps you will end up with a
>>> phased approach, starting with kettle and gradually folding the queries into
>>> a care2x module.  (Its quite a few years since I looked at care2x - I seem
>>> to recall php4 and smarty templates).
>>>
>>> Regards
>>> Bob
>>>
>>> On 23 May 2011 04:30, Juma Lungo <juma.lungo@xxxxxxxxxxxx> wrote:
>>>
>>>>  Dear Ola
>>>>
>>>> Can you take this case to the main DHIS2 developers' list. This case has
>>>> taken too long to be solved. As i shared before, the Care2x and DHIS2
>>>> integration tool  developed in Tanzania before, was not working at all.
>>>>
>>>> Lungo
>>>>
>>>> On 05/23/2011 01:14 AM, Maurizio Bricola wrote:
>>>>
>>>> Dear all,
>>>>
>>>>  after my visit in Arusha, I have tried to focus on how we could
>>>> proceed further especially with the MTUHA Report and Care2X.
>>>>
>>>>  I have read 3 times :) the report of Robert Meggle and I can say that
>>>> his analysis is correct, and the options he suggests are the right way to
>>>> proceed.
>>>>
>>>>  Indeed we have different options which are valid:
>>>>
>>>>  1) a little program in Java that can interface with the Care2X DB and
>>>> generates the XML file defining the right template. Every time we need to
>>>> import/export a report, we open the program and it will generate the file
>>>> that will be manually imported to the DHIS2 DB.
>>>> This program can be the one suggested by Robert
>>>> http://kettle.pentaho.com and will have the advantage of being there
>>>> already and have a graphic UI, which could be used also by non-programmers,
>>>> minimizing the realization time and making it easier to review the work.
>>>>
>>>>  2) use some scripts (php, java or even pearl etc.) which read the
>>>> Care2X DB and generate the XML file in the right format (compatible with
>>>> what DHIS2 is expecting to receive). After that the file will be manually
>>>> imported to DHIS2 DB. This will be basically a module (small application to
>>>> export data in xml compatible with what DHIS2 is expected to import)   of
>>>> Care2X, and will be composed of n files, among others a configuration file
>>>> where to specify the DB name, and other configuration like values conversion
>>>> etc. This will all depend on how the programmer will develop this module.
>>>>
>>>>  In less technical term the script will perform something like that:
>>>>
>>>>  Check DB xyz
>>>> Extract a, b, c, d
>>>> Convert into e, f, g, h
>>>> Generate XML
>>>> Save and/or download XML
>>>>
>>>>  If we opt for option 2 what we need is a team or a programmer with
>>>> knowledge of both systems to develop the module and knowledge of MTUHA
>>>> requirements.
>>>>
>>>>  With option one we need to know clearly the MTUHA requirements and
>>>> learn how to use the suggested software.
>>>>
>>>>  In order to decide it might be good if someone could have a look at
>>>> that software in details and run a test.
>>>>
>>>>  With Kind Regards,
>>>>
>>>> Maurizio Bricola
>>>> Technical Advisor (Kenya - Malawi - Tanzania)
>>>> International Institute for Communication and Development (IICD)
>>>> P.O. Box 11586, 2502 AN The Hague, the Netherlands
>>>> Visitors: Raamweg 5, 2596 HL The Hague
>>>> Phone: +31-(0)70-311 73 11 Mobile: +31-(0)6 33 77 35 41
>>>> Website: www.iicd.org
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>
>