Hey Peio, that sounds like an interesting project, please keep us updated as you go along! In late 2010 I tried porting OpenSpending (then WDMMG) over to RDF but couldn't make it work and decided the overhead was not justified by the expected gains [1]. I still think that for very fine-grained data (such as spending), you may be better off using an RDBMS or CSV (although I am increasingly a fan of putting URIs in both of those). What it gives you is ease of use, better aggregation and query language etc.
On the other hand, I like the idea of putting more "meta" data like actor profiles into a graph, although I've started doing a non-RDF version even here: https://github.com/pudo/grano - maybe a bit relevant? Have fun! - Friedrich [1] http://lists.okfn.org/pipermail/openspending/2010-November/000704.html On Wed, Feb 15, 2012 at 7:11 AM, Peio Popov <[email protected]> wrote: > Dear OKFN, > In the recent weeks I am playing and learning more about the general > idea of applying Semantic Technologies to the public information about > political funding and spending and the way it relates to the political > activity. > > My general idea of a project is to RDF-fy and link the open data about the: > > 1. Political parties, their budget subsidies, donations, annual > budgets, election campaign budgets and donations > 2. MP's linked to the parties, their profile, parliamentary activity > and annual tax declarations and > 3. Public officials and other data - "somehow" linked to the political > parties and the MPs > > If it sounds too broad and vague - it probably is. I am just reading > and experimenting for the fun of it. > > What I would like to achieve as a result is something that will enable > me to perform reasoning and inference based on the semantics of the > data. > > This kind of tasks are usually assisted by the application of an > ontology to the dataset so my first question is if you can recommend > such ontology in the field that I would generally call Linked > Political/Spending Data? > > I imagine further efforts in this direction would have to involve > web/text mining and applying semantic annotations according to the > conceptual models, so if you have any recommendations in this > directions - please share. > > Does any of these questions sounds like something you already met? > > I will be grateful for any response and general comment that might > help me in these wanders. > > Peio Popov > > _______________________________________________ > okfn-discuss mailing list > [email protected] > http://lists.okfn.org/mailman/listinfo/okfn-discuss _______________________________________________ okfn-discuss mailing list [email protected] http://lists.okfn.org/mailman/listinfo/okfn-discuss
