[wdmmg-dev] Mark's AidData

Mon Jun 20 08:55:03 UTC 2011

Hi Mark,

brilliant :-)

On Thu, Jun 16, 2011 at 5:04 PM, Mark Brough
<mark.brough at publishwhatyoufund.org> wrote:
> I installed OpenSpending a while ago on my laptop and it works OK after
> quite a bit of head-scratching - I think I pulled the Uganda data in - but
> a) it gave Error 500 when you clicked on a dataset (or spending or
> recipients)

Do you still have the error it gave around so you could post it here?

> b) I couldn't work out how to do visualisations
> so I got frustrated (mostly because I don't know Python) and gave up. In
> retrospect it was probably because the CSV mapping was wrong.

... And: do you have the CSV mapping? I've attached mine, which
relates to the pre-WB data model. It would need to be updated for the
newer IATI fields used in the WB data.

> I'll probably try and do a bit more work on it tonight, let me know when you
> want to have a chat about it.

I'll be online on Wednesday - are you around then?

> I think a good place to start would be: how difficult would it be for OS to
> pull IATI-XML data in directly itself? Or would we have to go via iati2csv?

Its very feasible to do this but requires writing a custom loader
script. We've done this for datasets in the past but it has generally
turned out to be a maintenance nightmare - so we're trying to keep
these two aspects seperate, hence iati2csv. I also think the CSV thing
lends itself to data fixing much better than straight IATI XML (i.e.
filling in sectors and regions) - also operations that you don't
necessarily want to perform in your loading script.

- Friedrich
-------------- next part --------------
A non-text attachment was scrubbed...
Name: model.json
Type: application/json
Size: 6586 bytes
Desc: not available
URL: <http://lists.okfn.org/pipermail/openspending-dev/attachments/20110620/28aaac03/attachment.json>