Due to recent changes in the GovInfo API, the Vistos GovInfo API components have become very unstable. Until a fix can be devised, the govinfo attributes on the Congress and CongressMember classes will be disabled.
Vistos is a module for downloading data on U. S. politicians from public sources via HTTP. V's goal is to empower U. S. citizens by providing an easier route for accessing the information necessary to hold public officials more accountable. This is to be accomplished by
- utilizing public data sources to gather information about public officials, and
- consolidating that information in a way that's easy to code around.
As programming isn't exactly a ubiquitous skill, it cannot go without saying that Vistos should only be considered a small, but fundamental, step in the far-greater goal of creating a more politically-informed populace. There is much more work to be done to fully realize such a goal, so Vistos merely seeks to provide the foundation for said work.
Interested? Jump to the samples below to learn how to set up Vistos.
pip install vistos
# your_project/script.py
import vistos as v
import pandas as pd
current_congress = v.Congress(116)
members = current_congress.bioguide.members
members_df = pd.DataFrame(members)
print(members_df.head())
>> python ~/your_project/script.py
bioguide_id first_name ... terms
0 S001165 Albio ... [{'congress_number': 109, 'term_start': 2005, ...
1 R000603 David ... [{'congress_number': 114, 'term_start': 2015, ...
2 S001172 Adrian ... [{'congress_number': 110, 'term_start': 2007, ...
3 C001049 William Lacy ... [{'congress_number': 107, 'term_start': 2001, ...
4 H001076 Margaret Wood ... [{'congress_number': 115, 'term_start': 2017, ...
Currently, the only public datasets supported by Vistos are the Biographical Directory of the United States Congress and the GovInfo API. This data can be downloaded in-bulk as tabular data using the Congress
object. More granular control can be achieved by using a CongressMember
object.
To use the GovInfo API, an API key is required. You can sign up for one here.
The CongressMember
class exists for querying data from the perspective of members. CongressMember
is a much faster option for when you know the specific member(s) you would like to download data for.
CongressMember
takes a Bioguide ID as an argument, and attempts to retrieve data for the member associated with the given ID.
The load
method manually loads the datasets specified when instantiating CongressMember
member = v.CongressMember('P000587', load_immediately=False)
member.load()
member_name = f'{member.first_name} {member.last_name}'
assert member_name == 'Mike Pence'
The bioguide
property returns Bioguide data as a BioguideMemberRecord
.
The govinfo
property returns GovInfo data as a dict
.
member = v.CongressMember('K000105', GOVINFO_API_KEY)
print(member.govinfo['title'])
Output:
Senator Edward M. Kennedy, Biography
Due to limitations in the GovInfo API, directly retrieving data about a member is not possible. Vistos attempts to work around these limitations by first requesting the members Bioguide data, and using the information found there to narrow down where to locate the member's data within the Congressional Directories of the Congresses the member belonged to.
The bioguide_id
property returns the selected Congress member's Bioguide ID.
The first_name
property returns the selected Congress member's first name.
The nickname
property returns the selected Congress member's nickname.
The last_name
property returns the selected Congress member's last name.
The suffix
property returns the suffix of the selected Congress member's name.
The birth_year
property returns the year that the selected Congress member was born.
The death_year
property returns the year that the selected Congress member died.
The biography
property returns biographical information about the selected Congress member.
The terms
property returns a list
of BioguideTermRecord
objects describing all of the terms the selected Congress member served.
Congress
is used to query a single congress, and takes either a year or number to determine which congress to return.
For example, the following Congress
objects all return the 116th U. S. Congress:
a = v.Congress(116)
b = v.Congress(2019)
c = v.Congress(2020)
assert a.bioguide == b.bioguide == c.bioguide
Excluding a year or number will return the active U. S. Congress:
c = v.Congress()
assert c.number == 116
Manually load datasets specified when instantiating Congress
c = v.Congress(116, load_immediately=False)
c.load()
Note: querying a transition year favors the congress that began that year (eg Congress(2015)
will return the 114th congress, not the 113th).
Calling get_member_bioguide()
returns a BioguideMemberRecord
corresponding to the given Bioguide ID.
Calling get_member_govinfo()
returns a dict
containing the GovInfo data corresponding to the given Bioguide ID.
The number
property returns an int
corresponding to the number of the selected Congress.
The start_year
property returns an int
corresponding to the first year of the selected Congress.
The end_year
property returns an int
corresponding to the last year of the selected Congress.
The bioguide
property returns Bioguide data as a BioguideCongressRecord
.
c = v.Congress(116)
print(c.bioguide)
Output:
{"members": [{ .. }], "congress_number": 116, "start_year": 2019, "end_year": 2021}
The govinfo
property returns GovInfo data as a GovInfoCongressRecord
.
The members
property returns a list
of unique CongressMember
objects:
c = v.Congress(116)
print(c.members[0].bioguide_id)
Output:
S001165
CongressBills
is a list
-based object for querying the bills for a single congress. This is in contrast to the Congress
object, which will attempt to download member data first.
congress_bills = v.CongressBills(105)
print(len(congress_bills))
Output:
13126
The load
method manually loads the dataset specified when instantiating CongressBills
search_bioguide_members(first_name: str, last_name: str, position: str, party: str, state: str, congress: int)
search_bioguide_members()
is a function for querying members by non-unique details.
When search_bioguide_members()
is called, queries will be sent as an HTTPS POST request to bioguideretro.congress.gov. The first_name
and last_name
parameters will match by the beginning of the string, but position
, party
, and state
will expect a selection from a discrete set of options. The available options can be found within the Party
, Position
, and State
classes found within the gpo
submodule.
search_govinfo_members(govinfo_api_key: str, first_name: str, last_name: str, position: str, party: str, state: str, congress: int)
search_govinfo_members()
works similarly to search_bioguide_members()
, but attempts to include GovInfo data for matching members.
The gpo
submodule is used within Vistos to perform basic tasks for retrieving data hosted by the United States Government Publishing Office. Within the gpo
submodule are helper classes and functions for creating more nimble scripts.
Returns the bioguide IDs of a given Congress number
Takes an input value and returns the corresponding congress number. If the number given is a valid congress number, it's returned as-is. If the number given is a valid year, the corresponding congress number is returned. Invalid postive numbers and None
return the current congress, and negative numbers return zero (the Continental Congress.)
Returns the number of the active congress, based on the current date
Returns the start year of the given congress number
Returns the end year of the given congress number
Returns a tuple containing the start and end years of the given congress number
Returns the congress numbers associated with a given year
Returns all congress numbers
A class containing options for the position parameter of search_congress_members()
A class containing options for the party parameter of search_congress_members()
A class containing options for the state parameter of search_congress_members()
An error for when an attempt is made to assign incorrectly-shaped data to the Congress.bioguide
or CongressMember.bioguide
properties.
An error for when an attempt is made to assign incorrectly-shaped data to the Congress.govinfo
or CongressMember.govinfo
properties.
An error for when an HTTP connection error is raised during Bioguide queries. This can be used to handle instability caused by requesting large amounts of members.
Below is an example of how to use the CongressMember
object and the gpo.get_bioguide_ids()
function to download all the members of the last ten Congresses, storing each in a JSON file per the first letter of their last name.
import os
import json
import shutil
import vistos as v
OUTPUT_DIR = './members_by_letter'
CURRENT_CONGRESS = v.gpo.get_current_congress_number()
def main():
# This script downloads data about Congress members for the past
# ten Congresses, using the CongressMember object in conjunction with
# the gpo.get_bioguide_ids() function.
# Get unique bioguides from the prior 10 years
all_bioguide_ids = set()
start_congress, end_congress = \
CURRENT_CONGRESS - 10, CURRENT_CONGRESS
for congress in range(start_congress, end_congress + 1):
bioguide_ids = v.gpo.get_bioguide_ids(congress)
all_bioguide_ids = all_bioguide_ids.union(bioguide_ids)
# Map Bioguide IDs to their corresponding letter
# of the alphabet (denoted by the first character)
members_by_alphabet = dict()
for bioguide_id in all_bioguide_ids:
letter = str(bioguide_id[0]).lower()
try:
member = v.CongressMember(bioguide_id, load_immediately=False)
members_by_alphabet[letter].append(member)
except KeyError:
members_by_alphabet[letter] = [member]
sorted_letters = sorted(list(members_by_alphabet.keys()),
key=lambda k: len(members_by_alphabet[k]),
reverse=True)
# For each letter, load and store the associated members
for letter in sorted_letters:
congress_members = members_by_alphabet.pop(letter)
member_headers = []
member_terms = []
while len(congress_members) > 0:
member = congress_members.pop(0)
bioguide_id = member.bioguide_id
# Download data
print(f'[{bioguide_id}] Downloading...', end='\r')
try:
member.load()
print(f'[{bioguide_id}] Downloaded ', end='')
except v.gpo.BioguideConnectionError:
print(f'[{bioguide_id}] Download Failed', end='')
finally:
print()
# Split terms from the rest of the data
member_header_record = dict(member.bioguide)
del member_header_record['terms']
member_headers.append(member_header_record)
member_terms.append({'bioguide_id': member.bioguide_id,
'terms': member.bioguide.terms})
# Save data
file_name = f'{start_congress}_{end_congress}_{letter}'
headers_path = create_path('headers', file_name)
terms_path = create_path('terms', file_name)
json.dump(member_headers, open(headers_path, 'w'))
print(f'[{letter.upper()}------] Saved Member data to {headers_path}')
json.dump(member_terms, open(terms_path, 'w'))
print(f'[{letter.upper()}------] Saved Term data to {terms_path}')
def pre_tasks():
if os.path.exists(OUTPUT_DIR):
shutil.rmtree(OUTPUT_DIR)
os.makedirs(OUTPUT_DIR)
def create_path(category, file_name):
if not os.path.exists(OUTPUT_DIR + '/' + category):
os.makedirs(OUTPUT_DIR + '/' + category)
return f'{OUTPUT_DIR}/{category}/{category}_{file_name}.json'
if __name__ == '__main__':
pre_tasks()
main()
Plainly stated, the function of Vistos is to enable people to more easily gather and present poltical information. This idea is meant to be the guiding thought for defining the scope of Vistos - that is to say that any data that enables U. S. citizens to be more politically informed can be considered an option for Vistos. To guide new additions to the project and prevent the project from falling into a state of over-ambitious aimlessness, all new work will be weighted by how easy it is to implement against what exists already.
At the moment, Vistos is a collection of Python-based classes that marry disparate data sources into more easily-managed objects. That doesn't mean that Vistos is inherently Python-based, or will never take another approach. Nor does it mean that it is poised to change anytime soon. It just means that a Python library currently makes the most sense for realizing the overall goal of Vistos, due to the popularity of Python and its ease of use. Ideally, Vistos is to stay in perpetual development and will always be taking the form of whatever makes the most sense at the time.
Currently, Vistos only supports Congressional data provided by the Government Publishing Office. Support for social media data and stocks are planned for implementation in the near future, after which, work on the components for the Executive branch will begin. On a more functional level, there are plans for verbosity and CLI support.
If you'd like to contribute to the project, or know of a useful data source, feel free to submit a pull request, or email z3c0.
-
Querying GovInfo does not return results for some congress persons.
When using the
CongressMember
class, some congress persons will not return GovInfo data. This is due to their GovInfo data missing a Bioguide ID, which is used for finding individual records within a Congressional Directory package. Should you need data for one of these members, the only workaround (currently) is to instead query GovInfo for the entire Congress that they served for, using theCongress
class. -
GovInfo data only goes as far back as the 105th Congress
The GovInfo API makes congress persons' data available via "Congressional Directories", which are only provided starting with the 105th Congress. If data for an earlier congress is needed, use Bioguide data instead.