[wdmmg-discuss] Working with OpenSpending platform from my perspective

Martin Keegan martin.keegan at okfn.org
Mon Aug 8 12:01:17 UTC 2011

On 8/4/11, Peio Popov <peio at peio.org> wrote:
> Dear list,
> In order to help data cartographer's education and make the
> introduction to Openspending platform easier, I would like to describe
> my experience with it.

Dear Peio,

thank you for all your constructive criticism, and perseverance.

> 1. Denormalized data. I spend more than an hour looking for a way to
> describe the nested levels in the raw data. I've always been told that
> I should normalize my data and if Rufus was not there to explain, I
> could never figured it out myself. It just seemed wrong to waste space
> like that. So please mention it when you speak about csv - it is not
> just a format to contain, but also a way to describe data.

Yes - we're aware that CSV (or denormalisation) isn't ideal.
Ultimately we'll provide alternative more appropriate means of loading

> 2. JSON. I did not feel very comfortable with it and it was good that
> I could use the model from the Israeli budget as a template. Looking
> at the model and then at the result helped me understand it better
> than the description in the wiki. It would be very helpful to have a
> JSON template with the main sections and an example of a typical
> record in it. I might propose one if you like the idea. Also any
> recommendation for an editor and syntax checker/highlighter is
> welcome.

Ok, that's a fair enough point - since you wrote this email, we've
massively updated our documentation in this area. Linking to official
examples (there's one on CKAN) is something we could still do better.

> 3. Decentralized info. Initially I was a bit confused between the OKFN
> sites. I wanted to achieve something like WDMMG, but I was told I need
> to look into Openspending, starting from CKAN. In the meantime I
> wandered to sites like Open Data Commons and Open Government data and
> OKFN site. I believe I found the wiki with google and there was the
> loading diagram. So please, link more to it - it is very helpful.

Yes, we also need to continue to consolidate our web estate. Most of
this effort has been reducing the number of systems used internally
rather than what we display externally.

> 4. Loading data. I used the irc based loading protocol and it worked
> good, but can be made better. The data wranglers should be able to
> help share this responsibility, because it is not always possible to
> interrupt your work to answer a question from a confused newbie.
> Thanks to borior and pudo's irc support I was able to not only load
> the data but also understand some of the platform concepts. I hope to
> contribute back and make their time with me well spend.

Thank you. At least the irc based loader works :)

> 5. Bubbles. As childish as it may sound I wanted the bubbles
> visualization from the very start. It was only because of pudo's hints
> and links that I could find the examples and the documentation. More
> links, more docs more functioning examples would be great. If I can do
> more visualizations, it would be even better.

Indeed, getting bubbles working out of the box is a key objective.

> 7. The result. As I've told you and wished on irc, this time i would
> not ask another stupid question, but will present the result - a
> visualisation of the Bulgarian budget spendings:
> http://peio.org/bgbudget/bubbles/bgbudget.html I hope to present it to
> some politicians in order to get promises for more data and perhaps
> even get the promises fulfilled.

Thank you very much for your collaboration so far. As I said, the docs
are now a LOT better:


... and your other comments (which basically everyone already agrees
with, I reckon :) ) will be taken on board.


More information about the openspending mailing list