How PlainInfluence Works, From Federal Data to Searchable Intelligence

Step 1: Data Collection

We download bulk data files directly from three federal systems. No scraping, no third-party intermediaries, we go straight to the source:

FEC bulk data - Committee master files, candidate records, and PAC-to-candidate contribution records. Downloaded as pipe-delimited flat files.
Senate LDA data - Lobbying registrations and quarterly activity reports. Downloaded as XML from the Senate system.
USAspending.gov - Contract award data from the Federal Procurement Data System. Downloaded via the USAspending API and bulk download portal.

Step 2: Entity Resolution

The core technical challenge is linking the same organization across three databases that use different naming conventions, different identifiers, and different data structures. Our entity resolution pipeline:

Name normalization - Standardize corporate suffixes (Inc, LLC, Corp), remove punctuation, handle common abbreviations.
FEC connected-organization matching - PAC registrations include a "connected organization" field that directly links the PAC to its corporate sponsor. This is the most reliable linkage.
Cross-database fuzzy matching - For organizations without explicit linkage, we apply name similarity algorithms to find potential matches across databases.
Parent-subsidiary resolution - Link subsidiaries and divisions to parent organizations where the relationship is clear from the data.
Conservative handling of ambiguity - Where a match is uncertain, records are kept separate rather than merged, so the pipeline never asserts a link it cannot support from the data.

Step 3: Aggregation and Enrichment

Once entities are linked, we aggregate the data to create useful profiles:

Organization profiles - Total PAC contributions to all candidates, total lobbying spending across all issue areas, total federal contract awards by agency.
Politician profiles - All PAC contributions received, broken down by contributing organization and amount.
Issue area profiles - Total lobbying spending by policy category, with organization breakdowns.
State profiles - Aggregated political money data for each state and territory.
Rankings - Top organizations by influence score, top politicians by PAC receipts, top lobbying spenders by issue.

Step 4: Serving the Data

The processed data is stored in a SQLite database and served through a server-side rendered web application. Every page is generated from the underlying data, there is no manual content creation for individual profiles. When the underlying data is updated, all profiles and rankings automatically reflect the changes.

What Makes This Different

Each of the three federal data sources is individually accessible. The FEC has a search interface. Senate LDA has a search portal. USAspending.gov has detailed contract lookup tools. What none of these systems provide is the cross-database linkage that shows the full picture: the same organization's PAC giving, lobbying spending, and government contract awards on a single page.

That linkage is what transforms three compliance databases into an intelligence platform. A journalist investigating defense industry influence can see, on one screen, how much a company spent lobbying the Department of Defense, how much its PAC gave to Armed Services Committee members, and how much it received in defense contracts. That synthesis would otherwise require hours of manual cross-referencing.

Accuracy and Limitations

We take data accuracy seriously. All monetary figures come directly from source filings - we never estimate, adjust, or project. Entity resolution is our primary source of potential error: an incorrect match could attribute one organization's activity to another. We prioritize precision (avoiding false matches) and flag ambiguous cases for review.

For detailed information about our data sources and their limitations, see our Data Sources page and Methodology.