This is an analysis from the Center for Public Integrity's data team in partnership with FiveThirtyEight. It was created by Chris Zubak-Skees. The analysis was informed by questions posed by political reporter Carrie Levine.
- Analysis - An analysis of ActBlue and other FEC filings.
- Data - Spreadsheets produced by the analysis.
What does it mean when one organization raises more than $2.9 billion for democrats? The Center set out to answer that question.
This analysis relied on a PostgreSQL database of more than 289,500 campaign finance filings loaded by the Center using fec-loader. These were analyzed using SQL, Pandas and Jupyter notebook software.
The notebook was developed on MacOS 10.13 with git, pipenv and Python 3 installed. These instructions are specific to that environment, but should be adaptable to others:
git clone https://github.com/PublicI/actblue-analysis.git
cd actblue-analysis
pipenv install
pipenv shell
python -m ipykernel install --user --name=actblue-analysis
jupyter labQuery results are persisted to Pickle files, so the notebook can be run without access to the database.
If you do have access to a PostgreSQL database of campaign finance records in the correct schema and want to modify the queries, copy config.json.example to config.json, then update the values in config.json to the correct hostname and password.