hateimparlament/vignettes/funwithdata.Rmd

---
title: "funwithdata"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{funwithdata}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

```{r setup}
library(hateimparlament)
library(dplyr)
library(ggplot2)
library(stringr)
library(tidyr)
```

## Preparation of data

First, you need to download all records of the current legislative period.
```r
fetch_all("../inst/records/") # path to directory where records should be stored
```
Second, those `.xml` files, need to be parsed into `R` `tibbles`. This is accomplished by:
```r
read_all("../inst/records/") %>% repair() -> res
```
We also used `repair` to fix a bunch of formatting issues in the records and unpacked
the result into more descriptive variables.

For development purposes, we load the tables from csv files.
```{r}
res <- read_from_csv('../inst/csv/')
```
and unpack our tibbles
```{r}
comments <- res$comments
speeches <- res$speeches
speaker <- res$speaker
talks <- res$talks
applause <- res$applause
```

## Analysis

Now we can start analysing our parsed dataset, e.g. find out which party gives the most talks:
```{r, fig.width=7}
join_speaker(res$speeches, res) %>%
    group_by(fraction) %>%
    summarize(n = n()) %>%
    arrange(n) %>%
    bar_plot_fractions(title="Number of speeches given by fraction",
                        ylab="Number of speeches")
```

or counting the occurences of a given word:

```{r, fig.width=7}
find_word(res, "Kohleausstieg") %>%
    filter(occurences > 0) %>%
    join_speaker(res) %>%
    select(content, fraction) %>%
    filter(!is.na(fraction)) %>%
    group_by(fraction) %>%
    summarize(n = n()) %>%
    arrange(desc(n)) %>%
    bar_plot_fractions(title = "Parties using the word 'Kohleausstieg' the most (absolutely)",
                        ylab = "Number of uses of 'Kohleausstieg'",
                        flipped = F)
```

### Who gives the most speeches?

```{r}
res$speeches %>%
    group_by(speaker) %>%
    summarize(n = n()) %>%
    arrange(-n) %>%
    left_join(res$speaker, by=c("speaker" = "id")) %>%
    head(10)
```

### Who talks the longest?

```{r}
res$talks %>%
    mutate(content_len = str_length(content)) %>%
    group_by(speaker) %>%
    summarize(avg_content_len = mean(content_len)) %>%
    arrange(-avg_content_len) %>%
    left_join(res$speaker, by=c("speaker" = "id")) %>%
    head(10)
```

### Which party gives the most applause to which parties?

```{r}
res$applause %>%
    left_join(res$speaker, by=c("on_speaker" = "id")) %>%
    select(on_fraction = fraction, where(is.logical)) %>%
    group_by(on_fraction) %>%
    arrange(on_fraction) %>%
    summarize("AfD" = sum(`AfD`),
              "BÜNDNIS 90 / DIE GRÜNEN" = sum(`BUENDNIS_90_DIE_GRUENEN`),
              "CDU/CSU" = sum(`CDU_CSU`),
              "DIE LINKE" = sum(`DIE_LINKE`),
              "FDP" = sum(`FDP`),
              "SPD" = sum(`SPD`)) -> tb
```

For plotting our results we reorganize them a bit and produce a bar plot:

```{r, fig.width=7}
pivot_longer(tb, where(is.numeric), "by_fraction", "count") %>%
    filter(!is.na(on_fraction)) %>%
    bar_plot_fractions(x_variable = on_fraction,
                        y_variable = value,
                        fill = by_fraction,
                        title = "Number of rounds of applauses from fractions to fractions",
                        xlab = "Applauded fraction",
                        ylab = "Rounds of applauses",
                        filllab = "Applauding fraction",
                        flipped = FALSE)
```


### Which party comments the most on which parties?

```{r}
res$comments %>%
    left_join(res$speaker, by=c("on_speaker" = "id")) %>%
    select(by_fraction = fraction.x, on_fraction = fraction.y) %>%
    group_by(on_fraction) %>%
    summarize(`AfD` = sum(str_detect(by_fraction, "AfD"), na.rm=T),
              `BÜNDNIS 90 / DIE GRÜNEN` = sum(str_detect(by_fraction, "BÜNDNIS 90/DIE GRÜNEN"), na.rm=T),
              `CDU/CSU` = sum(str_detect(by_fraction, "CDU/CSU"), na.rm = T),
              `DIE LINKE` = sum(str_detect(by_fraction, "DIE LINKE"), na.rm=T),
              `FDP` = sum(str_detect(by_fraction, "FDP"), na.rm=T),
              `SPD` = sum(str_detect(by_fraction, "SPD"), na.rm=T)) -> tb
```
Analogously we plot the results:

```{r, fig.width=7}
pivot_longer(tb, where(is.numeric), "by_fraction", "count") %>%
    filter(!is.na(on_fraction)) %>%
    bar_plot_fractions(x_variable = on_fraction,
                        y_variable = value,
                        fill = by_fraction,
                        title = "Number of comments from fractions to fractions",
                        xlab = "Commented fraction",
                        ylab = "Number of comments",
                        filllab = "Commenting fraction",
                        flipped = FALSE)
```

### When are which topics discussed the most?

```{r, fig.width=7}
pandemic_pattern <- "(?i)virus|corona|covid|lockdown"
climate_pattern <- "(?i)klimawandel|erderwärmung|co2|treibhaus|methan|kyoto-protokoll|klimaabkommen"
pension_pattern <- "(?i)rente|pension|altersarmut"

word_usage_by_date(res, c(pandemic = pandemic_pattern,
                          climate = climate_pattern,
                          pension = pension_pattern)) %>%
    ggplot(aes(x = date, y = count, color = pattern)) +
    xlab("date of session") +
    ylab("occurence of word per session") +
    labs(color = "Topic") +
    geom_point()
```