Skip to content

Master plan workshop notes

Jump to bottom
cyberbikepunk edited this page Jul 3, 2016 · 1 revision

July 1 2016 Workshop

People

  • Balazs Hungary
    • follow the money
    • front-end POC underway
  • Christian
    • Getting the data verified
  • Tim
    • Starting his bachelor in computing at TU
    • interested in data science and ML
  • Anna did wonderfully
  • Benjamin Came in late
  • Monisha is Canadian. Canadians are not like Americans. It's Canada day today. Yeah.
  • Paul Ozzie
  • Loic with umlaut

Status Quo

  • list of leads and hunches
  • list of interested journalists
  • frontend hungary
  • Grant applications
    • ESJ
    • Google DNF pilot pending
    • bigger fundds
  • Data
    • England scraped
    • FOIA Berlin
    • Denmark
    • Github repo: 10 countries
  • Spatial data
    • what level of aggregation?
    • we have it to individual transaction by law
    • some countries release the bare minimum, some more
    • lots of advocacy to do / outreach
    • FOIA + lawyer
  • Pilots
    • go and find interested journalists
    • start with countries with data
      • poland
      • hungary
      • germany
      • croatia
    • newspapers
      • local papers
      • need to have regional audience
      • go for the pilot story for the pilot

Post-it time

  • data that can be understood by the average Joe
  • training + tuturials for working with cohesion data
  • political affiliattion
  • all data geooacted
  • transparency as a solution
  • media coverage
  • brexit and uk subsudies
  • list of eu subs per capita
  • map w/ drill down
  • scalable thing
  • front end for journalists
  • dashboard
  • verify data
  • I want to believe in the EU
  • Impact
  • I want OS to provide a good generic backend for data analysis and public API for custum UIs (anonymous ozzie)
  • categorize transactions
  • A sweet and savoury scandal => news
  • A systems perspective on money flow => book

4 clusters of ideas

  • impact (ultimate philosophical goals)
  • data goals + backend
  • interface/front-end
  • stories

Random notes:

Combine data with other data

  • electoral data
  • per capita
  • sociology is difficult (going below the national level is difficult)

Reverse thinking: Cohesion funds to draw in attention to the UX

Anonymous Ozzie: journalists don't use dashboards, they ask specific questions.

Cross comparison: It would be nice to be able to compare what's comparable. i.e. this small town in Portugal vs this small town in Poland are in similar situations, why don't they get the same money dude? (technical mental note: link datasets). How much does 1km of motorway cost here and there?

Technical mental note "background datasets" for a lack of a better word.

Trainings for journalists: win win situations. we get feedback from the local level and we help them how use the data to dig further into the story.

Mental post-it: OS backend should be able to query all datasets for local insights. i.e I'm here lat:lon, what are the money transactions around me?

Specialist to generalists: Journalists should drill down on specific stories, refering the users to the more general dataset.

Co-design definition: build you technical stack with the user through the process of iteration. Let's face it: we don't do it.

Crowd sourced verification: it's not easy do do! How can you really tell if the money was well spend?. The money disapears in places that are not directly apparent, i.e. travels budgets etc...

Whistle blowing: is the upstream equivalent of crowd-sourcing verication, which happens downstream. It's more accurate. But it's a different ball game.

Anonymous dutch farmer: we need to be defining a very basic set of generic requirements for data access/interface and build custom apps for local stories according to demand (I'm paraphrasing) ...mmm... comes dangerously close to OS top-level goals...

Our data is not bullshit in the first place. Do we really need to verify it before releasing it? It's the journalist's job to do it as he or she is writing the story. But we can help them do it. And when the data is questionnable, that's actually a story in itself.

Crowd sourcing addresses: could help correct geolocating companies.

The "how dare you ask" question does OS work with translations?

Wrap-up

Assigning tasks: ohoh i'm gonna get some homework

Bjorn Borg: we're doing the same thing in parallel. Let's coordinate.

Data collection: i knew i was gonna get some homework.... quote: we're happy to do that! (my boss)

Pilot: let's start working on the Berlin data! Get extra support data. (which might be a really long process). CHR (pronounced XXXXRRRRR) doesn't see the story behind the data for Berlin yet.

Real wrap-up

  • Local stories
    • what data do you need for a local story exactly? Find out the hard way by trying to write an article
    • proof-of-concept article?
    • OhOh. Someone just pulled a deadline out out of the hat! Berlin/German story by August?
  • Brexit story (cause it's just cool)
    • that's cleary a local story
    • the fact that you can't get a hold of the data is the story
  • Formatted FOIA request for the media to use (FragdenStadt style)
    • FOAI attack!
  • Data collection + backend:
    • Bela&Anna to finish gathering meta-data
    • get a point of contact in countries for requesting any missing data (make a list of countries concerned: Anna by monday 6am)
    • then everything into next.openspending
    • the rest is history

Tomorrow

Cleaning and modelling the data. Sounds like fun? Come join us. Thank you Anna!!!!!!!!

Clone this wiki locally