In Nicolas-Schmidt/speech: Legislative Speeches

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "80%"
)

The speech package

Nicolás Schmidt, Diego Luján, Juan Andrés Moraes

r badger::badge_devel("Nicolas-Schmidt/speech", "orange")

Description

Converts the floor speeches of Uruguayan legislators, extracted from the parliamentary minutes, to tidy data.frame where each observation is the intervention of a single legislator.

Installation

# Install speech from CRAN
install.packages("speech")

# The development version from GitHub:
if (!require("remotes")) install.packages("remotes")
remotes::install_github("Nicolas-Schmidt/speech")

Data generation process

1 - Floor speeches

2 - Data extraction

3 - First construction of the data set: speech::speech_build()

4 - Final data set: speech::speech_build(., compiler = TRUE)

Example

You can see more complex examples in the following link.

library(speech)
url <- "https://parlamento.gub.uy/documentosyleyes/documentos/diarios-de-sesion/6084/IMG"
text <- speech::speech_build(file = url)
text


speech_check(text, initial = c("A", "M"))


text <- speech::speech_build(file = url, compiler = TRUE)
text


text$word <- speech_word_count(text$speech)

dplyr::glimpse(text)

Possible application

library(magrittr)

minchar <- function(string, min = 3){
    string <- stringr::str_remove_all(string, "[[:punct:]]")
    string <- unlist(strsplit(string, " "))
    string[nchar(string) > min]
}

text$speech %>% 
    minchar(., min = 4) %>%  
    quanteda::corpus() %>% 
    quanteda::dfm(remove = c("señor", "señora")) %>% 
    quanteda.textplots::textplot_wordcloud(color = rev(RColorBrewer::brewer.pal(10, "RdBu")))

library(ggplot2)

text$speech %>% 
    minchar(., min = 4) %>%  
    tibble::enframe() %>% 
    tidytext::unnest_tokens(word, value) %>%
    dplyr::count(word, sort = TRUE) %>%
    dplyr::mutate(word = stats::reorder(word, n)) %>%
    dplyr::filter(!stringr::str_detect(word, "^señor")) %>% 
    .[1:40,] %>% 
    ggplot(aes(word, n)) +
        geom_col(col = "black", fill = "#00A08A", width = .7) +
        labs(x = "", y = "") +
        coord_flip() +
        theme_minimal()

Detecting roll-call votes in parliamentary speeches

urls <- speech_url(chamber = "D", days = c("2002-06-12", "2004-04-14"))
rollcall <- speech_rollcall(file =  urls)

rollcall

summary(rollcall)

Citation

To cite packagespeech in publications, please use:

citation(package = 'speech')

Maintainer

Nicolas Schmidt ([e-mail])

Nicolas-Schmidt/speech documentation built on July 4, 2023, 4:32 p.m.