- Background
- Project Ideas
- assessing the Stimulus
- identifying personae and scenarios for ARRA data use
- tracking end-to-end flow of money and data
- Goals and benchmarks for ARRA
- Contextualizing government agencies and relate to overall federal spending
- Making the ARRA data more friendly to developers
- Tying together media coverage
- Engaging social media
Background
Official US sources
-
The US government's official site for the public:
Recovery.gov
-
The website for ARRA recipients to report in:
Home Page - FederalReporting.gov
-
You can download the recipient data from
Download Center
Other Projects to track the stimulus
-
Eye on the Stimulus - ProPublica -- the best reporting I've seen on ARRA. See
How Much Stimulus Funding is Going to Your County? UPDATED Dec. 2009 | ProPublica Recovery Tracker
-
Fedspending.org: Recovery Act Recipient Reports Database --
Awards performed in Pennsylvania 14 Congressional District
-
Stimulus Watch: Keeping an Eye on Economic Recovery Spending -- e.g.,
15275 | PA | Performance Place | StimulusWatch.org
-
a partial (but editable!) list at
Stimulus and bailout web projects - Sunlabs wiki
-
Sunlight Labs: Blog - Recovery.gov Augmented Reality Mashup -- an iPhone app!
Our work at Berkeley
-
Stimulus collab with Erik Wilde and Eric Kansa:
Proposed Guideline Clarifications for American Recovery and Reinvestment Act of 2009 and
Web Services for Recovery.gov
Project Ideas
Preliminary thoughts on how I plan to teach my Berkeley course with a focus on the Stimulus:
Making Sense of the Stimulus
assessing the Stimulus
Some big questions:
-
Is the stimulus working?
-
Did it accomplish what was promised?
-
Should there be a follow-up to the Stimulus?
-
How do we help various constituencies make sense of the stimulus for their purposes?
-
Is there a fraud, waste, or abuse?
identifying personae and scenarios for ARRA data use
A list I brainstormed of the types of people possibly interested in ARRA data
-
journalists
-
the Recovery Board
-
politicians
-
job-seekers
-
government agency people at federal, state, and local levels
-
watchdog groups
-
technologists
-
ordinary citizens
-
open data advocates
-
people looking to get a piece of the stimulus pie: grants, loans, or contracts
-
political strategists, consultants: Democrats versus Republicans versus independents
-
academics looking to write the history of ARRA and compare that to the great depression
-
economists do research the Recovery
-
citizen journalists
-
local journalists
-
"citizen inspector generals"
-
government transparency advocates
How do figure out who we really want to serve and who is really interested?
tracking end-to-end flow of money and data
I want to make clear the flow of data so that others can reproduce the results put out by the government. Transparency should include reproducibility.
For example, there are tons of numbers on recovery.gov. Can we download the raw numbers and reproduce what we see on the website? Can we show clearly how the final numbers are reached?
The conceptual and monetary flow (roughly):
pre-legislation wrangling -> final bill -> Treasury Department -> agencies -> states, contractors -> subcontractors, vendors = results vs metrics
Let's map out where the pieces are.
Goals and benchmarks for ARRA
Nailing down the overall goals for ARRA and how they have been broken down to the smallest levels is very helpful. For instance, on recovery.gov, we can track individual contractors and how much money they have spent, and what program that money came from -- but right now we can't easily go all the way from the legislation that initially authorized that program and see how that program was put into place. The goal here is to make that flow very easy to understand.
What are the high level goals of ARRA?
The link http://www.whitehouse.gov/assets/Documents/Recovery_Plan_Metrics_Report_508.pdf is now dead. Fortunately, someone had uploaded it to http://www.scribd.com/doc/11333017/Recovery-Plan-Metrics-Report-508
http://www.recovery.gov/?q=content/act used to have a bunch of overall goals listed -- link now broken... The goals were:
-
Save and create more than 3.5 million jobs over the next two years;
-
Take a big step toward computerizing Americans’ health records, reducing medical errors, and saving billions in health care costs;
-
Revive the renewable energy industry and provide the capital over the next three years to eventually double domestic renewable energy capacity;
-
Undertake the largest weatherization program in history by modernizing 75 percent of federal building space and more than one million homes;
-
Increase college affordability for seven million students by funding the shortfall in Pell Grants, increasing the maximum award level by $500, and providing a new higher education tax cut to nearly four million students;
-
As part of the $150 billion investment in new infrastructure, enact the largest increase in funding of our nation’s roads, bridges, and mass transit systems since the creation of the national highway system in the 1950s;
-
Provide an $800 Making Work Pay tax credit for 129 million working households, and cut taxes for the families of millions of children through an expansion of the Child Tax Credit;
-
Require unprecedented levels of transparency, oversight, and accountability.
Contextualizing government agencies and relate to overall federal spending
-
make an org chart of the US so that we can have blurbs for the departments, agencies, and bureaus.
-
it turns out that the US budget database is a good source of data of the structure of US government.
-
make learning about the federal (or state or local) government into a game.
-
write-up of where I left off:
My project idea for the Freebase Hack Day and the org chart:
An org chart of the US Federal Government Based on OMB agency and bureau codes
-
U.S. Government Manual: Main Page is another good source: it's in PDF -- and maybe we can be a more structured format -- but if we can't, we should turn the Gov't Manual into a database and structured document.
Making the ARRA data more friendly to developers
-
recast into a relational form and possibly redistribute in that form (sqlite? MySQL?). I've been experimenting with an Access DB and can indeed verify that there are 130362 awards = 13080 grants + 607 loans + 11667 contracts -- if you filter out the vendors. Note that the spreadsheets have 157142 awards in total = 143352 grants + 710 loans + 13080 contracts (without the filtering of vendors).
-
upload the data into Freebase to create guids for all entities (recipients, agencies, contracts, etc.) There are unique identifiers for awards but I've not found easy ways to link to any of them via those identifiers.
-
wrap the data into an API. (There is no API currently available from recovery.gov.)
Tying together media coverage
Reporters are writing about ARRA all around the country about topics big and small. How can you assemble that so that you can zoom into any part of the country and see coverage about those issues? Crawl the news feeds and try to tie to specific projects and locations?
Engaging social media
2010 is the year of the mid-term elections. How can Stimulus data be used (misused?) in the political battles that will ensue around job creation and ARRA? See, for example, this twitter feed:
Twitter / Search - recovery.gov
[....more ideas to come ....]
