---
title: "Intro to the unvotes package"
author: "David Robinson"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{The unvotes package}
  %\VignetteEncoding{UTF-8}
  %\VignetteEngine{knitr::rmarkdown}
editor_options: 
  chunk_output_type: console
---

```{r echo = FALSE}
knitr::opts_chunk$set(message = FALSE, warning = FALSE,
                      fig.height = 5, fig.width = 5)
```

This package provides the voting history of countries in the [United Nations General Assembly](http://www.un.org/en/ga/), along with information such as date, description, and topics for each vote.

### Datasets

The unvotes contains three datasets, each data frames (specifically tbl_dfs, which are more convenient to print). First is the history of each country's vote. These are represented in the `un_votes` dataset, with one row for each country/vote pair:

```{r}
library(dplyr)
library(unvotes)

un_votes
```

The package also contains a dataset of information about each roll call vote, including the date, description, and relevant resolution that was voted on:

```{r}
un_roll_calls
```

Finally, the `un_roll_call_issues` dataset shows relationships betwen each vote and 6 issues:

```{r}
un_roll_call_issues

library(dplyr)
count(un_roll_call_issues, issue, sort = TRUE)
```

(Use `help()` to get information and documentation about each dataset).

### Example analysis

Many useful analyses will first involve joining the vote and roll call datasets by the shared `rcid` (roll call ID) column:

```{r joined}
library(dplyr)

joined <- un_votes %>%
  inner_join(un_roll_calls, by = "rcid")

joined
```

One could then count how often each country votes "yes" on a resolution in each year:

```{r by_country_year, dependson = "joined"}
library(lubridate)

by_country_year <- joined %>%
  group_by(year = year(date), country) %>%
  summarize(votes = n(),
            percent_yes = mean(vote == "yes"))

by_country_year
```

After which this can be visualized for one or more countries:

```{r by_country_year_plot, dependson = "by_country_year"}
library(ggplot2)
theme_set(theme_bw())

countries <- c("United States", "United Kingdom", "India", "France")

# there were fewer votes in 2013
by_country_year %>%
  filter(country %in% countries, year <= 2013) %>%
  ggplot(aes(year, percent_yes, color = country)) +
  geom_line() +
  ylab("% of votes that are 'Yes'")
```

Similarly, we could look at how the voting record of the United States has changed on each of the issues by joining with the `un_roll_call_issues` dataset:

```{r issue_plot, dependson = "joined", fig.height = 8, fig.width = 8}
joined %>%
  filter(country == "United States") %>%
  inner_join(un_roll_call_issues, by = "rcid") %>%
  group_by(year = year(date), issue) %>%
  summarize(votes = n(),
            percent_yes = mean(vote == "yes")) %>%
  filter(votes > 5) %>%
  ggplot(aes(year, percent_yes)) +
  geom_point() +
  geom_smooth(se = FALSE) +
  facet_wrap(~ issue)
```