Projects/Where Does My Money Go/Notes

Done

 * Parse + clean up PESA data from HMT
 * See http://www.openeconomics.net/wdmmg/
 * First try at basic visualisation
 * See http://www.openeconomics.net/wdmmg/dept

Priority

 * Start a Google doc and draft the plan/letter to go to cabinet office (jwyg)
 * Continue to extract/clean up PESA data (rgrp) - see p/Government_Finances/PESA
 * Document specific ideas for different visualisations (lizturner)

Backburner

 * Look at open visualisation packages on: http://wiki.okfn.org/OpenVisualisation
 * Continue to investigate whether there is more data that could be easily incorporated

Visualisation ideas

 * Treemap with timeline representing HMT PESA data
 * See also p/Where_Does_My_Money_Go/Visualizations

2009-06-23
Highlights of IRC discussion 2009-06-23

#!irc if you haven't looked already:  this already listed a good set of tables from PESA rgrp: this looks quite detailed - but is there any geo information? so pesa chapter 09 we have: Table 9.1 Total identifiable expenditure on services by country and region, 2002-03 to 2007-08 yeh, i was thinking - if you could drill down into a tree map, and have some little motion chars embedded inside... okay i've really got to load this stuff up -- i've written the code i just haven't hooked it up yet ... ephidrina: do you think it would be useful to have a list of different visualisation ideas somewhere? but yeh, inside each function think about some little customized graphics which make sense for that dept i'm also thinking of autoloading all the spreadsheets into google docs as a backend editing facility so that this can be interactive ... (but that is separate from viz stuff!) yes, certainly a list of ideas is not bad even better is if we can match the model to the data stream i.e. think about what context is best for which quantity rgrp: is that pesa ch9 online somewhere? explain? yes http://www.hm-treasury.gov.uk/pespub_pesa08.htm that's where I currently scrape from and then tidy up and put online what i meant was: there are many different models we can use to visualize stuff but we have to identify the most meaningful way to model different things like - when to use a tree map or a pie chart, for instance right, yes the hard problem is getting the model right both from a data extraction and a visualization pov exactly ultimately my guess is we do want something like a traditional accounting model with transactions and accounts we can then output from that to the format needed by visualization that sounds like a good plan would it be complicated to overlay geographical data onto a model like that? but that is a lot of work! yes and no the problem is this: if you want to display a given facet in your output (visualizations, etc)_ you need it represented in your account structure e.g. rather than just having unemployment expenditure you'd need this broken down into each region that you wanted to use ah, so you have to make sure you do it from day 1 yes -- or do work later on splitting a given account into subaccounts the one way round is just have stuff tagged e.g. we tag this expenditure as being on unemployment and the north-east problem then is reconciliation -- if we tag a given expenditure as being both, say, NE and SW w/o indication of how expenditure is split we get double-counting etc this is one of problems with PESA data: it is already sliced up for presentational purposes and getting back to original accounts is hard ... my feeling atm is to go for something that just works i.e. we look at the data, decide what we can visualize and we simply do some hacking to get the code out for each case on the next iteration we can worry about getting the perfect model ... yeh, i think getting a nice, meaningful presentation with the resoutrces we have is the priority exactly how do you feel about using the motion charts, btw? i guess my feeling here is we work with whatever tools will give the best results in the time available looking at the PESA data i think we probably have enough to be going on with i think they could give us some nice bang for the buck, but i am not sure how flexible they will be i guess we could see how things go do something simple to start with and work up no harm in trying, i guess :) exactly :) our top priority right now is to actually put in for the money from cabinet office ah rght - what do you need for that? i think we need to send them a brief doc outlining what we need money for and who we plan to get to do the work perhaps the doc should contain a short list of requirements from the public's perspective yes i think we could take our existing outline like - who this is for, and what will they want? http://wiki.okfn.org/p/Where_Does_My_Money_Go/ and what sort of a presentation are we going to make to reach them? yes, like that, but a bit more concrete i agree though we should probably acknowledge for prototype route that we are going to be driven by the data availability than by exactly what we want one thing i thought would be nice is a basic top down overview of spending (from pesa) with a time slider - e.g. as a pie chart i wonder how easy it would be to get a timeline slider onto the treemap.... i don't think it would be that hard what would happen to all of the parts when you moved the slider? i started looking into animation js stuff for: http://www.rufuspollock.org/2009/05/22/visualizing-technology-flows-over-time-i/ anyways ... b4 we go off-track it would be nice if you could show aggregated spending over a time interval... plus i think we could dump down these visualization ideas *with* the data they would need to make them happen so we need to define the scope of teh project inputs and outcomes a prototype, and a presenation? audience is supposed to be the general citizen who wishes to understand these issues better and how do we reach them? my guess is through the web and perhaps media coverage hm - it kind of complicates things to think about publicity at this stage so i guess let's just think about what's reallistic with existing resources existing/projected exactly ok, so now where? 1. Let's start a google doc and draft the plan/letter to go to cabinet office 2. let's dump the visualization ideas + data requirements on wiki page 3. I'll work on extracting the rest of the PESA data 4. We can start looking through this data and seeing what match we have with 2 5. once money is definitely confiremed we start working ... how does this sound? that sounds very sensible we'll make some sugggestions in 2. and take a look at scope for design/development ok so immediate actions are 1 + 2 with main responsibility for 1 being jwyg and me (with some input from you guys) main responsibility for 2 (no more than 20m of brain dumping) is with you ugys started wiki page at http://wiki.okfn.org/p/Where_Does_My_Money_Go/Notes ok - we can pull some useful material form the discussion there ah great i think what's coming out of the mailing list is very interesting some nice examples of what works and doesn't work yeah might be nice to populate the notes page a bit then post back to the list for more ideas! ok we should probably decide quite quickly what we're going to do i think the data tail wagging the visualization dog should be our mantra ;) ok i agree so that's 3+4 jwyg: are you happy to take on creating google doc and drafting (with my assistance) yup sure i'll get PESA data out asap shall we aim to meet up again next Tuesday afternoon to discuss progress? yeh, we'll pitch in on google doc sure next tue we'll aim to have a draft *before* next tuesday so we can sign it off (one hopes) then maybe even sooner i'll be travelling next tuesday.. back on wednesday - is that possible instead? yeh, wed is actually better for us... ok, let's do wednesday jwyg: you may want to mark next wednesday as an OKF collab day instead of/as well as tuesday ok - i need to split now will stay on the # in case something else comes up nice talking to you - looking forward to an interesting outcome! let us know when google doc is in place great, thanks for dropping in liz speak soon bye liz!