This is a set of libraries and command line utilities for parsing information from Congress, the Sunlight Foundation, and others. The purpose is to track the influence of money in politics.
Wiki
Information about the project, api documentation, and data documentation
can be found at the project wiki.
The wiki serves as the knowledge repository of the project. Feel free to add or elaborate
on the information contained therein.
Components
Below is information about the various components of the utilities.
Legislator Events
This does two things: 1) creates a json object with info about all of the actions taken by, or events relating to, a given legislator and 2) visualizes those actions and events on an interactive timeline.
The CLI is inspired by https://github.com/unitedstates/congress.
Setup
cd legislator-events
cp config.example.py config.py # add your sunlight api key to config.py
To see the list of available senators:
./run list-legislators --chamber=senate
Available representatives can be listed by passing 'house' to the chamber flag. Both chambers will be listed if one or the other is not specified.
To see the list of available events/actions to include in the output object:
./run list-events
To generate the json for a given legislator:
./run legislator-events --legislator="John A. Boehner"
This command will create a data folder to store the output json (and subsequent output files) as well as a cached folder for reusable data files that aren't output files themselves.
Note: the legislator names are the full official names listed by the list-legislators task. This is usually First Last, but sometimes names deviate.
To see the timeline:
cd legislator-events/timeline
python -m SimpleHTTPServer 8000 # d3 xhr cross origin requests are only supported for HTTP.
Open a browser and got to http://localhost:8000
Legislation (legis)
This trolls through legislative xml trying to find identical paragraphs of bills. Outputs a json file which contains a dictionary whose keys are the sha1 of the text of the paragraphs of the legislation fed in and whose values are lists of appearance in the order (name_of_bill, paragraph_id, full_text_of_paragraph).
To use this you will need to get the full text of the bills that congress passed. You can either scrape this yourself from Thomas (which is painful) OR you can get it from GovTrack.us. They like to use the rsync protocol, so make sure you have rsync, and then do
mkdir -p bill_directory/congress_number
rsync -avz --delete --delete-excluded govtrack.us::govtrackdata/us/congress_number/bills.text
For example, if you want to compare the 110th through 112th congresses, you would first rsync as follows:
for c in {110, 111, 112};
do
mkdir -p bill_directory/${c};
rsync -avz --delete --delete-excluded govtrack.us::govtrackdata/us/${c}/bills.text;
done;
Note the sha is computed from text of a paragraph stripped of the lowercased text of a paragraph stripped of all numbers, symbols and subtags. Hence, the full text of paragraphs may differ significantly. For example, you will quickly see that there is a very large number of pargraphs which read like
For the fiscal year 2012, $10,000,0000.
but have different years and values. This is an artifact of the fact that Congress sets the budget and that this is the standard phrasing for budget bills and I sha only forthefiscalyear.