Skip to content
Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

voteview member_photos repo

Build Status

This repository contains photos of U.S. congressional representatives through the ages, as well as code necessary to regenerate this data from scratch. We currently have approximately 10,175 of 12,475 representatives accounted for including every member serving since 1945.

This project is a part of, a website dedicated to providing information about historical U.S. legislators, including NOMINATE ideological scores, historical roll-call votes, and biographical information. is a project of University of California Los Angeles' Department of Political Science. The corresponding maintainer for this repository is Aaron Rudkin.

How to Use

Using the images

members.csv contains a list of all members we have photos for at the time the file was generated. This will allow you to map familiar names to the ICPSR IDs that index our photo filenames. The file is sorted by most recent congress served, then alphabetically. The photos presented are scaled to 600px in height, 4x5 aspect ratio. Files smaller than 600px in height are not upscaled, and images very near 4x5 aspect ratio are not cropped.

Example results:

Name ICPSR State Party Congress Chamber Born Died Image Source Provenance
WELLSTONE, Paul David 049101 Minnesota Democratic Party 107 Senate 1944 2002 images/bio_guide/049101.jpg bio_guide
CLINTON, William Jefferson (Bill) 099909 President Democratic Party 106 President 1946 images/wiki/099909.jpg wiki
GUILL, Ben Hugh 003874 Texas Republican Party 81 House 1909 1994 images/manual/003874.jpg manual Representing Texas

Installing prerequisites to add images

In order to add images, you will need to install several dependencies. The scrapers, which seek and download new images, require Python and several libraries. The image processing side, which resizes and crops images, requires ImageMagick, smartcrop-cli, JPEGTran, and jpegoptim.

To install external dependencies, users running environments with support for apt or brew can run

To install Python dependencies, use poetry install via Poetry or pip install -r requirements.txt.

Quick tutorial: adding new images

  1. Run a scraper or manually add a photo to the appropriate raw folder (likely images/raw/manual/).
  2. If a manual image has been added, add a provenance statement to config/provenance.json
  3. Run to generate processed versions of the images from raw images.
  4. Run config/ to update the database with the new images. This ensures that the members.csv file is up to date and that's tests work.
  5. Run to ensure the data integrity.
  6. If added images upgrade earlier images (for example, bio_guide images replacing wiki images), run --flush to remove the no longer used files.
  7. Open a pull request to submit your images to us.

Check for Missing allows users to check for representatives whose photos are missing and generates a table based on criteria provided.


  • --type flat: Use a flatfile database instead of our default MongoDB instance. Most end users should use this argument.
  • --min N: Provide a number N which represents the minimum Congress to scan for missing photos (default 81 [1947-1949])
  • --max N: Provide a number N which represents the maximum Congress to scan for missing photos. Default is left black.
  • --chamber chamber: Province a chamber chamber describing a specific chamber of congress. Valid options are House or Senate. Default is left blank.
  • --state state: Province a two-character state postal abbreviation to limit searches to one state. Example: CO for Colorado.
  • --sort sort: Provide a string sort which describes which field to sort on. Valid options are bioname, icpsr, state_abbrev, party_code, congress. Default is congress. When grouping, you can also sort by Amount.
  • --year: If specified, table will include "year" instead of "congress" and the --min and --max arguments will expect a year.
  • --raw: If specified, the script will check for images where we have processed copies, but no raw copies. Clones of the repository that have not yet re-scraped the raw files from bio_guide and wiki should see all such images; clones of the repository that have scraped images should report no missing raw files.
  • --group [state_abbrev | congress]: If specified, instead of printing a table of individual missing images, a count grouped by the group parameter will be printed. Useful to see which states or congresses are complete.

Example usage:

python --type flat --min 50 --state CT --chamber House --sort bioname

Scrape Congressional Bioguide allows users to scrape the Congressional Bioguide for photos.


  • --type flat: Use a flatfile database instead of our default MongoDB instance. Most end users should use this argument.
  • --min N: Provide a number N which represents the minimum Congress to scan for missing photos (default 20)

Example usage:

python --type flat --min 50

Scrape Wikipedia allows users to scrape Wikipedia for photos.


  • --type flat: Use a flatfile database instead of our default MongoDB instance. Most end users should use this argument.
  • --min N: Provide a number N which represents the minimum Congress to scan for missing photos (default 20)
  • --icpsr ICPSR --url "http://...": Provide an ICPSR and a URL to manually scrape a Wikipedia article for that ICPSR. Useful when the default name or search is inadequate. The resulting page will still be checked against the scoring algorithm to ensure the page is appropriate for the member.
  • --override 1: By default, we cache data from Wikipedia articles so that we don't check for every congressperson every time we run. Use this argument to override the cached data and re-scrape every users who would otherwise fit the parameters. Useful during a cutover of congress.
  • --blacklist ICPSR: Mutually exclusive to all other arguments; tells the scraper to not scrape this ICPSR from Wikipedia in the future. Useful when the correct page has a photo that is incorrectly scraped (i.e. house or memorial photo or military insignia instead of photo of person).

Example usage:

python --type flat --min 50 will scrape photos for all our currently known cases where the default scraper scrapes an incorrect photo or misses the search query.

Manual Photos

Some photos were collected manually from other sources. In addition to distributing the already-resized versions of these, raw versions of these photos (best available quality/resolution) are stored in images/raw/manual/. Information about where each of these images came from is stored in config/provenance.json. These images are automatically downsampled and cropped when running the processing steps below.

Facial Recognition

In order to ensure a more uniform set of images, it may be desireable to ensure that all images face the same direction. We use Azure for a facial recognition API. If API keys are contained in config/facial_recognition.json, then (which resizes and re-aspects input images) will additionally detect which direction the image is facing and if necessary flip it so that it is facing stage left (our right). Code describing the lookup is in under needs_horizontal_flip().

To set up this API, copy config/facial_recognition_blank.json to config/facial_recognition.json and fill out the two fields with valid credentials.

Process Photos

  • will resize, re-aspect, flip, and optimize images. Images will move from images/raw/<source>/<file>.<ext> to images/<source>/<file>.jpg.
  • will scrape Bioguide, Wikipedia, perform the manual Wikipedia overrides, and then constrain the images in order. This should generate the repository essentially as-is from scratch.


  • config/config.json: User-Agent for scraper and some default URLs, as well as database connection info if you are connecting to a MongoDB database to search members.
  • config/facial_recognition_blank.json: A blank template for inserting Azure Face API key/endpoint, see section Facial Recognition for details
  • config/bio_guide_results.json: Blacklist for Congressional bioguide.
  • config/wiki_results.json: Blacklist for Wikipedia and greylist (articles recently scraped, confirmed to contain nothing, skip for a while)
  • config/parties.json: Party metadata, used for both checking Wikipedia articles and outputting party names.
  • config/states.json: State metadata, used for both checking Wikipedia articles and outputting party names.
  • config/database-raw.json: Large raw database dump, used for flat-file searches. Generated by config/

Behind the Scenes:

  • config/ Dumps current Mongo database to flatfile (use this first, to update the local flat file). Requires our local MongoDB instance.
  • config/ Dumps the current images to a members.csv file. Can take --type flat to dump from flat file. (Use this after updates to correctly log the information in the menbers file)
  • Runs basic sanity tests to ensure data is running correctly. Used in our travis-CI build.
  • Uploads the current folder's raw images to our S3 store.
  • Downloads our S3 store's set of raw images to your local copy of this repository.
  • Powers some of the image resizing behind the scenes.


We welcome contributions of photos or code improvements. For code improvements, please open a pull request.

For sources for photos, please see our Issues page. If you are contributing a photo to an existing project, just reply with a comment including the photo (highest resolution possible, include information about where the photo is from and any rights issues). If no project seems applicable, or if you are letting us know about a new source of many photos, please open a new Issue. We believed that the use of low resolution images of historical public figures, freely obtained largely from public domain or government sources, constitutes fair use. Please ensure that any images you suggest are cleared for use by and users of this repository.


Photos of U.S. congressional representatives and scrapers used to collect them.




No releases published


No packages published