EDA Chamber decisions.Rmd

---
title: "EDA Chamber Decisions GFCC"
author: "Sebastian Sternberg"
date: "2 September 2019"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```



Load data and packages

```{r}
rm(list = ls())

require(magrittr)
require(dplyr)
require(ggplot2)
require(lubridate)
require(stringi)
require(stringr)
require(strex) #regex etc.
require(ggthemes)

#load data

load("chamber_decision_cleaned.Rda")

df <- chamber_decisions

#fill NAs of kammer_txt (with kammer of citation)

df$kammer_txt[is.na(df$kammer_txt)] <- df$kammer_cit[is.na(df$kammer_txt)]

#some NAs in the AZ:

df$link <- as.character(df$link)

#We extract them from the link:

df$az[is.na(df$az)] <- str_extract(df$link[is.na(df$az)], "\\dbvr\\d+")

#take care of NAs in Senat. Senat is always the first number of the AZ:

df$senat[is.na(df$senat)] <- str_first_number(df$az[is.na(df$senat)])

#no more NAs!
sapply(df, function(y) sum(length(which(is.na(y)))))

```


# Summarize by date

### Summarize by year, month, week day

```{r}

df$year <- year(df$date)
df$month <- month(df$date) %>% as.factor()
df$weekday <- wday(df$date, label = T)

```


```{r}


df <- df %>% filter(year >1997)

df_year <- df %>% group_by(year) %>% tally()
df_year_senat <- df %>% group_by(year, senat) %>% tally()
df_year_senat_chamber <- df %>% group_by(year, senat, kammer_txt) %>% tally()



#in one call
ggplot(tally(group_by(df, year, senat)),
    aes(x = year, y = n, fill = senat)) +
    geom_bar(stat="identity") + labs(fill="Employment")


ggplot(df_year, aes(x = year, y = n)) +
    geom_bar(stat="identity")


#year by Senate
ggplot(df_year_senat, aes(x = year, y = n, fill = senat)) +
    geom_bar(stat="identity") + labs(fill="Senate")



#year by Senate chamber


count_pct <- function(df) {
  return(
    df %>%
      tally %>% 
      mutate(n_pct = 100*n/sum(n))
  )
}


df_senat1 <- df %>% filter(senat == 1) %>% group_by(year, kammer_txt) %>% count_pct()
df_senat2 <- df %>% filter(senat == 2) %>% group_by(year, kammer_txt) %>% count_pct()


outfolder <- "C:/Users/ssternbe/Dropbox/Kaggle/Scraper_GFCC_Decisions/scraper-decisions-German-Federal-Constitutional-Court/out/figures/"

pdf(paste0(outfolder, "/","perc_Senate1byChamber.pdf"))

ggplot(df_senat1, aes(x = year, y=n_pct, fill=kammer_txt)) + 
    geom_bar(position="fill", stat="identity") + labs(title = "Senate 1", fill="Chamber") + 
  theme_tufte()
dev.off()

pdf(paste0(outfolder, "/","perc_Senate2byChamber.pdf"))


ggplot(df_senat2, aes(x = year, y=n_pct, fill=kammer_txt)) + 
    geom_bar(position="fill", stat="identity") + labs(title = "Senate 2", fill = "Chamber")+
  theme_tufte()


dev.off()

```




per weekday


```{r}

#aggregate by weekday

ggplot(tally(group_by(df, weekday)),
    aes(x = weekday, y = n)) +
    geom_bar(stat="identity") + theme_tufte()


ggplot(tally(group_by(df, senat, weekday)),
    aes(x = weekday, y = n, fill = senat)) +
    geom_bar(stat="identity") + labs(fill="Senate")



```

per month:

```{r}

ggplot(tally(group_by(df, month)),
    aes(x = month, y = n)) +
    geom_bar(stat="identity") + theme_tufte()




```