dumbbells.Rmd

---
title: "2020 NZ Election"
output: 
  flexdashboard::flex_dashboard:
    theme: spacelab
    self_contained: false
    css: https://fonts.googleapis.com/css?family=Open%20Sans
    lib_dir: lib
    mathjax: null
    navbar:
      - { title: "Intro", icon: "fa-home", href: "index.html", align: left }
      - { title: "Parties", icon: "ion-ios-settings-strong", align: left }
      - { title: "Refs", icon: "ion-ios-settings-strong", href: "referenda.html", align: left }
      - { title: "Hexmaps", icon: "ion-cube", href: "hexmaps.html", align: left }
      - { title: "@dakvid", icon: "fa-twitter", href: "https://twitter.com/dakvid", align: right }
    social: menu
    source_code: https://github.com/dakvid/election2020
editor_options: 
  chunk_output_type: console
---
  
  ```{r init, include=FALSE}
knitr::opts_chunk$set(echo = FALSE)
knitr::opts_chunk$set(message = FALSE)
knitr::opts_chunk$set(warning = FALSE)
```

```{r setup, include=FALSE}
library(readr)
library(dplyr)
library(forcats)
library(glue)
library(stringr)
library(plotly)

GE2020 <- readRDS("data/GE2020.rds")
GE2017 <- readRDS("data/GE2017.rds")
electorates <- read_csv("data/electorates.csv")
party_colours <- read_csv("data/party_colours.csv")

GE2017 <- 
  GE2017 %>% 
  inner_join(
    electorates %>% select(id, id_14),
    by = "id_14"
  ) %>% 
  select(-id_14, -electorate_name_14)

party_voters_2017 <-
  GE2017 %>% 
  filter(VotingType == "Party",
         !str_detect(Party, "nformal")) %>% 
  group_by(id) %>% 
  summarise(Voters2017 = sum(Votes)) %>% 
  ungroup()

candidate_voters_2017 <-
  GE2017 %>% 
  filter(VotingType == "Candidate",
         !str_detect(Candidate, "nformal")) %>% 
  group_by(id) %>% 
  summarise(Voters2017 = sum(Votes)) %>% 
  ungroup()

party_voters_2020 <-
  GE2020 %>% 
  filter(VotingType == "Party",
         !str_detect(Party, "nformal")) %>% 
  group_by(id) %>% 
  summarise(Voters = sum(Votes)) %>% 
  ungroup()

candidate_voters_2020 <-
  GE2020 %>% 
  filter(VotingType == "Candidate",
         !str_detect(Candidate, "nformal")) %>% 
  group_by(id) %>% 
  summarise(Voters = sum(Votes)) %>% 
  ungroup()
```


# 

```{r db_data}
party_votes <-
  GE2020 %>% 
  filter(VotingType == "Party") %>% 
  group_by(id, Party) %>% 
  summarise(Votes = sum(Votes)) %>% 
  ungroup() %>% 
  inner_join(party_voters_2020, by = "id") %>% 
  mutate(PerCent = round(Votes / Voters * 100, 2))

party_votes_2017 <-
  GE2017 %>% 
  filter(VotingType == "Party") %>% 
  group_by(id, Party) %>% 
  summarise(Votes2017 = sum(Votes)) %>% 
  ungroup() %>% 
  inner_join(party_voters_2017, by = "id") %>% 
  mutate(PerCent2017 = round(Votes2017 / Voters2017 * 100, 2))

pv_change <-
  party_votes %>% 
  full_join(party_votes_2017,
            by = c("id", "Party")) %>% 
  inner_join(electorates, by = "id") %>% 
  select(id,
         Electorate = electorate_name,
         Electorate14 = electorate_name_14,
         boundary_change,
         Party,
         `2020` = PerCent,
         `2017` = PerCent2017)
```


## Plots {.tabset}

```{r db_fn}
show_dumbbell <- function(db_party_name = NA_character_, round_up = TRUE) {

  the_change <- 
    pv_change %>% 
    filter(Party == db_party_name)
  if (round_up) {
    the_change <- 
      the_change %>% 
      mutate(`2020` = round(`2020`, 1),
             `2017` = round(`2017`, 1))
  }
  
  the_change %>% 
    mutate(Electorate = fct_reorder(Electorate, `2020`)) %>% 
    inner_join(party_colours, by = "Party") %>% 
    plot_ly(color = I("gray80")) %>% 
    add_segments(
      x = ~pmin(`2017`, `2020`), xend = ~pmax(`2017`, `2020`),
      y = ~Electorate, yend = ~Electorate,
      showlegend = FALSE
    ) %>% 
    add_markers(
      x = ~`2017`, y = ~Electorate, name = "2017",
      color = ~if_else(boundary_change, I("grey90"), I("grey60")),
      hoverinfo = "text",
      text = ~glue("{`2017`}% of {Electorate14} votes in 2017 {if_else(boundary_change, '<br><i>Boundary differs from 2020</i>', '')}")
    ) %>% 
    add_markers(
      x = ~`2020`, y = ~Electorate, name = "2020",
      color = ~I(party_colour),
      hoverinfo = "text",
      text = ~glue("{`2020`}% of {Electorate} votes in 2020 {if_else(is.na(`2017`), '<br><i>New electorate</i>', if_else(boundary_change, '<br><i>Boundary changed</i>', ''))}")
    ) %>% 
    layout(
      title = glue("Relative Votes for the {db_party_name} by Electorate"),
      xaxis = list(title = "Votes (Percent)"),
      yaxis = list(title = "", dtick = 1),
      margin = list(l = 130),
      legend = list(x = 0.8, y = 0.1)
    )
}
```

### Labour

```{r db_labour}
show_dumbbell("Labour Party")
```

### National

```{r db_national}
show_dumbbell("National Party")
```

### Green

```{r db_green}
show_dumbbell("Green Party")
```

### ACT

```{r db_act}
show_dumbbell("ACT New Zealand")
```

### Māori

```{r db_māori}
show_dumbbell("Māori Party", round_up = FALSE)
```

### NZ First

```{r db_nzf}
show_dumbbell("New Zealand First Party")
```

### TOP

```{r db_top}
show_dumbbell("The Opportunities Party (TOP)", round_up = FALSE)
```

### Cannabis

```{r db_aclp}
show_dumbbell("Aotearoa Legalise Cannabis Party", round_up = FALSE)
```

### Left vs Right

```{r db_left_right}
pv_change_lr <-
  pv_change %>% 
  mutate(Group = 
           case_when(Party %in% c("Labour Party", "Green Party", "Māori Party") ~ "Left",
                     Party %in% c("National Party", "ACT New Zealand", "New Conservative", "Conservative") ~ "Right")) %>%
  filter(!is.na(Group)) %>%
  group_by(Group, Electorate, Electorate14, boundary_change) %>%
  summarise(`2020` = sum(`2020`, na.rm = TRUE),
            `2017` = sum(`2017`, na.rm = TRUE)) %>%
  ungroup() %>% 
  # deal with new electorate
  mutate(`2020` = if_else(`2020` == 0, NA_real_, round(`2020`, 1)),
         `2017` = if_else(`2017` == 0, NA_real_, round(`2017`, 1)))
pv_change_lr <-
  inner_join(
    pv_change_lr %>% 
      filter(Group == "Left") %>% 
      select(Electorate, Electorate14, boundary_change, `Left 2020` = `2020`, `Left 2017` = `2017`),
    pv_change_lr %>% 
      filter(Group == "Right") %>% 
      select(Electorate, Electorate14, boundary_change, `Right 2020` = `2020`, `Right 2017` = `2017`)
  ) %>% 
  mutate(Electorate = fct_reorder(Electorate, `Left 2020`))
plot_ly(pv_change_lr, color = I("gray80")) %>% 
  add_segments(
    x = ~pmin(`Left 2017`, `Left 2020`), xend = ~pmax(`Left 2017`, `Left 2020`),
    y = ~Electorate, yend = ~Electorate,
    showlegend = FALSE
  ) %>% 
  add_segments(
    x = ~pmin(`Right 2017`, `Right 2020`), xend = ~pmax(`Right 2017`, `Right 2020`),
    y = ~Electorate, yend = ~Electorate,
    showlegend = FALSE
  ) %>% 
  add_markers(
    x = ~`Left 2017`, y = ~Electorate, name = "Left 2017",
    color = ~if_else(boundary_change, I("grey90"), I("grey60")),
    hoverinfo = "text",
    text = ~glue("{`Left 2017`}% of {Electorate14} voted for Lab/Grn/Māo in 2017 {if_else(boundary_change, '<br><i>Boundary differs from 2020</i>', '')}")
  ) %>% 
  add_markers(
    x = ~`Right 2017`, y = ~Electorate, name = "Right 2017",
    color = ~if_else(boundary_change, I("grey90"), I("grey60")),
    hoverinfo = "text",
    text = ~glue("{`Right 2017`}% of {Electorate14} voted for Nat/ACT/Cons in 2017 {if_else(boundary_change, '<br><i>Boundary differs from 2020</i>', '')}")
  ) %>% 
  add_markers(
    x = ~`Left 2020`, y = ~Electorate, name = "Left 2020",
    color = I("red1"),
    hoverinfo = "text",
    text = ~glue("{`Left 2020`}% of {Electorate} voted for Lab/Grn/Māo in 2020 {if_else(is.na(`Left 2017`), '<br><i>New electorate</i>', if_else(boundary_change, '<br><i>Boundary changed</i>', ''))}")
  ) %>% 
  add_markers(
    x = ~`Right 2020`, y = ~Electorate, name = "Right 2020",
    color = I("blue1"),
    hoverinfo = "text",
    text = ~glue("{`Right 2020`}% of {Electorate} voted for Nat/ACT/NCons in 2020 {if_else(is.na(`Right 2017`), '<br><i>New electorate</i>', if_else(boundary_change, '<br><i>Boundary changed</i>', ''))}")
  ) %>% 
  layout(
    title = '"Left" and "Right" Votes by Electorate',
    xaxis = list(title = "Votes (Percent)", dtick = 10),
    yaxis = list(title = "", dtick = 1),
    margin = list(l = 130)
  )
```




## Explanation {.sidebar}

These dumbbell plots show the percent of the party vote received in each electorate,
highlighting the shift from 2017, with the the caveat that there were boundary changes
to around half of the electorates. Darker grey dots had the same boundaries and lighter
grey have had changes.

As the headlines have been shouting, Labour had universal increases and National had
universal decreases.
ACT too had universal increases (though it would have been hard to have gone down).

The Greens had big increases in their strongholds and modest increases in most other places.
I think the couple of notable decreases are due to boundary changes.

The Māori Party overall went down, which is unsurprising after three years out of
parliament. It is interesting to see that the Māori electorates saw both jumps and
drops - notably with less support in Waiariki, which was their only candidate win.

NZ First universally plummeted, being portrayed as Jacinda's "hand brake".
TOP couldn't make as much of a splash as they did last time, with drops
across the board.

Would the cannabis referendum enthuse or dampen support for the Legalise Cannabis
Party? They did indeed see a boost, most strikingly across the Māori electorates.

Finally, how did the vote share change between "the Left" (Labour, Green and Māori)
vs "the Right" (National, ACT, Conservative and New Conservative)?
The nationwide shift left
[begun in 2017](https://david.frigge.nz/election2017/dumbbells.html#left-vs-right)
continues, with only three electorates remaining with a right-wing majority,
and a further eight with a right-wing plurality.



<script>
  (function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
  (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),
  m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
  })(window,document,'script','https://www.google-analytics.com/analytics.js','ga');

  ga('create', 'UA-30013353-2', 'auto');
  ga('send', 'pageview');

</script>