Want to automate campaign finance data gathering from the Federal Election Commission? The FEC’s OpenFEC API is flexible from a developer’s perspective and cumbersome from an analyst’s. tidyusafec is a data retrieval R package that anticipates a desire to work with tidy campaign finance data directly in R.
Functions cover searching for elections, candidates, and committees and then getting more detailed data from their filings. While still very much in development, the most popular endpoints are covered and the package strategy reasonably stable. Functions can optionally return the raw results in list format. A tidy example below:
# Load R packages library(tidyusafec) library(tidyverse) # Get Data candidates <- search_candidates(office = 'H', election_year = '2018', district = '05', state = 'VA', unnest_committees = FALSE) candidate_totals <- candidates %>% get_candidate_totals() # Wrangle Data candidate_totals %>% filter(type_of_funds %in% tidyusafec_filters$candidate_totals$type_of_funds$receipts_smallest_components, cycle == '2018') %>% arrange(-desc(amount)) %>% mutate(type_of_funds = type_of_funds %>% str_replace_all('_',' ') %>% str_to_title()) %>% ggplot() + # Visualize Data geom_bar(aes(x = forcats::as_factor(name), y = amount, fill = party), stat = 'identity') + facet_wrap(~type_of_funds, labeller = label_wrap_gen(width = 10)) + scale_y_continuous(labels = scales::dollar) + coord_flip() + theme_bw() + scale_fill_manual(values = c('DEM' = '#2c4d82', 'REP' = '#8e1b1b', 'IND' = '#a3a3a3', 'GRE' = '#1c561d', 'LIB' = '#afac3d', 'NNE' = '#0A0A0A')) + labs( title = 'Virginia Congressional District 05 Race - 2018', subtitle = 'Financial Summary', x = '', y = '', fill = 'Party', caption = paste0('Source: FEC. Data retreived on ', Sys.Date(), ', created by @StephenHolz') ) + theme(axis.text.x = element_text(angle = -90, hjust = 0, vjust = .5))
A range of examples of the library in action is over in the explore section, covering all House and Senate races of the past 10 years. These are meant to be seeds for faster reproducible analysis and will typically need extra work to be cleaned up for a larger audience.
The package documentation site can help you get started :)