vignettes/acts.md

In titaniumtroop/aRlegislation: A Dataset of Arkansas Legislation from 2001-2019

Acts

This vignette examines the structure of the acts tibbles contained within the legislation dataset and displays the partisanship of the Arkansas Legislature over time as an exmample use case.

library(aRlegislation)
library(dplyr)
library(tidyr) # needed for nest/unnest operations
library(ggplot2)

The acts tibbles contain the text of laws passed during each cycle/session, along with a little bit of metadata about the acts:

head(legislation$acts[[1]])
#> # A tibble: 6 x 6
#>     act partisan  bill_no title             subtitle            text            
#>   <dbl> <chr>     <chr>   <chr>             <chr>               <chr>           
#> 1     1 committee 1       AN ACT TO MAKE A… AN ACT FOR THE ARK… "SECTION 1. APP…
#> 2     2 D         1001    AN ACT TO MAKE A… AN ACT FOR THE ARK… "SECTION 1. APP…
#> 3     3 bipartis… 1143    AN ACT TO CHANGE… AN ACT TO CHANGE T… "SECTION 1.    …
#> 4     4 committee 1156    AN ACT TO MAKE A… THE GENERAL APPROP… "SECTION 1. APP…
#> 5     5 D         68      AN ACT TO PROVID… TO PROVIDE THAT AN… "SECTION 1.    …
#> 6     6 committee 1168    AN ACT TO MAKE A… AN ACT FOR THE ARK… "SECTION 1. APP…

The act text has been lightly processed such that it is still human-readable. A really short act provides an example:

SECTION 1. Arkansas Code 14-121-305(b)(2), pertaining to drainage improvement districts, is amended to read as follows: (2) No work exceeding ten thousand dollars ($10,000) twenty thousand dollars ($20,000) shall be let without public advertisement.

The act titles, subtitles, and text are suitable for natural language processing analysis, although examples of such are beyond the scope of this introduction. The text above illustrates one caveat: for acts that are amended, both the deleted text and the new text appears in the corpus [e.g., the act referenced above increases the dollar threshold from ~~ten thousand dollars ($10,000)~~ to twenty thousand dollars ($20,000)].

The metadata includes a calculated partisan value based on the political party of the bill sponsors, as follows:

• If any sponsor is a committee, the partisan value is set to committee
• If all sponsors are lawmakers in the same party, the partisan value is set to the value of that party
• If the sponsors are lawmakers of mixed political parties, the partisan value is set to bipartisan

The partisan value allows us to look at trends about bipartisanship over time, as follows:

party.colors <- c(
  "R" = "#990000", # dark red = Republicans
  "D" = "#668cff", # light blue = Democrats
  "G" = "#00cc00", # bright green = Green party
  "I" = "#444444", # dark grey = Independent
  "unk" = "#b3b300", # dark yellow = unknown
  "bipartisan" = "#8A2BE2", # purple
  "committee" = "#888888" # medium grey
)

legislation %>%
  unnest(acts) %>%
  filter(partisan == "bipartisan") %>%
  count(cycle, partisan) %>% # counts number of parties sponsoring acts
  ggplot(aes(y = n, x = cycle, color = partisan)) +
    geom_point(size = 2.5) +
    geom_line(size = 3) +
    geom_smooth(method = "lm", color = "blue", se = T) +
    scale_x_continuous(
      breaks = seq(from = 2001, to = 2019, by = 4), 
      minor_breaks = seq(from = 2001, to = 2019, by = 2)
    ) +
    scale_color_manual(values = party.colors) +
    labs(
      title = "Total Bipartisan Acts",
      x = "",
      y = "",
      caption = ""
    ) +
    theme(
      legend.text = element_text(size = 6),
      legend.position = "none"
    )
#> `geom_smooth()` using formula 'y ~ x'

We can see here that the total number of bipartisan acts has been declining over time.

titaniumtroop/aRlegislation documentation built on May 4, 2020, 3:24 a.m.