hexmaps.Rmd

---
title: "2017 New Zealand Election"
output: 
  flexdashboard::flex_dashboard:
    theme: spacelab
    self_contained: false
    lib_dir: lib
    mathjax: null
    navbar:
      - { title: "Intro", icon: "fa-home", href: "index.html", align: left }
      - { title: "Dumbbells", icon: "ion-ios-settings-strong", href: "dumbbells.html", align: left }
      - { title: "Hexmaps", icon: "ion-cube", href: "", align: left }
      - { title: "3D Hexmaps", icon: "fa-cubes", href: "3dhex.html", align: left }
      - { title: "@dakvid", icon: "fa-twitter", href: "https://twitter.com/dakvid", align: right }
    social: menu
    source_code: https://github.com/dakvid/election2017
---

```{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(flexdashboard)
library(sf)
library(magrittr)
library(dplyr)
library(forcats)
library(purrr)
library(stringr)
library(leaflet)
library(leaflet.minicharts)
library(nzelect)
library(htmltools)
library(htmlwidgets)

begin_label_text <- "<div style='font-size:12px;float:left'>"
begin_label_text_200 <- "<div style='font-size:12px;width:200px;float:left'>"
begin_label_header <- "<span style='font-size:18px;font-weight:bold'>"
end_label_header <- "</span><br>"
end_label_text <- "</div>"

load("data_geo/nzhex.rda")
load("data_votes/GE2017.rda")
data("GE2014")
GE2014 %<>%
  mutate(id = Electorate %>% str_extract("[0-9]+") %>% as.integer(),
         Party = if_else(Party == "Maori Party", "Māori Party", Party))

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

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

party_voters_2014 <-
  GE2014 %>% 
  filter(VotingType == "Party",
         !str_detect(Party, "nformal")) %>% 
  group_by(id) %>% 
  summarise(Voters2014 = sum(Votes)) %>% 
  ungroup()

candidate_voters_2014 <-
  GE2014 %>% 
  filter(VotingType == "Candidate",
         !str_detect(Candidate, "nformal")) %>% 
  group_by(id) %>% 
  summarise(Voters2014 = sum(Votes)) %>% 
  ungroup()

getLeafletOptions <- function(minZoom, maxZoom, ...) {
  leafletOptions(
    crs = leafletCRS("L.CRS.Simple"),
    minZoom = minZoom, maxZoom = maxZoom,
    dragging = TRUE, zoomControl = FALSE,
    tap = FALSE,
    attributionControl = FALSE , ...)
}

getFactorPal <- function(f) {
  colorFactor(colormap::colormap(
    colormap = colormap::colormaps$hsv,
    nshades = length(f)), f)
}
```

# {.storyboard}

### Hello HexMaps: a hexagonal cartogram of New Zealand electorates (2014-2017)

```{r hex_regions}
leaflet(options = getLeafletOptions(5.8, 5.8)) %>% 
  addPolygons(
    data = nzhex,
    label = ~nzhex %>% 
      transpose() %>% 
      map(~ paste0(begin_label_text,
               begin_label_header, .x$electorate_name, end_label_header,
               "In ", .x$region, ".",
               end_label_text) %>% HTML()),
    weight=2, color='#000000', group = 'Regions',
    fillOpacity = 0.6, opacity = 1, fillColor = ~getFactorPal(region)(region),
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addLabelOnlyMarkers(
    data = nzhex %>% st_centroid(),
    label = ~abbr,
    labelOptions = labelOptions(
      noHide = 'T', textOnly = T,
      offset=c(-6,-11))
  ) %>% 
  onRender(
    "function(el, t) {
      var myMap = this;
      // get rid of the ugly grey background
      myMap._container.style['background'] = '#ffffff';
      myMap.scrollWheelZoom.disable();
    }"
  )
```

***

This is a [cartogram](https://en.wikipedia.org/wiki/Cartogram) of the current
(2014-2017) New Zealand electorates, with one hexagon for each electorate.
Electorates have roughly equal populations, so this is a reasonable map of
population distribution. The North and South Islands should be obvious; the
Māori electorates make up a second mini NZ in the bottom right.

Why are these maps more useful than a geographic map? In short, because they don't
overrepresent rural areas and underrepresent urban areas.
[Stephen Beban (NZ)](https://thespinoff.co.nz/politics/24-09-2017/a-better-visual-breakdown-of-the-2017-election-results/) and
[Pitch Interactive (USA)](http://pitchinteractive.com/latest/tilegrams-more-human-maps/)
have more to say on the topic.

The colours here indicate the regional groupings, to help orient you to the
distorted geography. Many electorates cross regional boundaries, so these
are just a rough guide. Most notably, the Wellington region is split between
several Māori electorates. (Credit for the regional groupings to
[Chris Knox in the Herald](http://insights.nzherald.co.nz/election/)).

This hexagonal layout is due to 
[Joseph Wright](http://mapdruid.blogspot.co.nz/2015/05/hexagonal-tile-map-of-new-zealand.html),
and I used his shapefiles as a base.

Chris McDowall made [a different hexagonal layout](http://web.archive.org/web/20140307180437/http://hindsight.clerestories.com/2014/01/06/chris-mcdowall-hexagonal-maps/)
for the 2011 electorates, which [I adapted](http://david.frigge.nz/nzhex2014/) in 2014.
I think I like that better in terms of the geographic layout, but it is very
skinny so it's less compact for on-screen display.

Click through the storyboard and see different data from the election
visualised in hexmaps...


### Electorate Winners: Who won the seats?

```{r hex_candidates}
electorate_winners <-
  GE2017 %>%
  filter(VotingType == "Candidate") %>%
  group_by(id, Party) %>%
  summarise(Votes = sum(Votes)) %>%
  top_n(1, Votes) %>%
  ungroup() %>% 
  inner_join(candidate_voters_2017) %>% 
  mutate(PerCent = round(Votes / Voters * 100, 1)) %>% 
  mutate(party_colour = case_when(
    Party == "National Party" ~ parties_v["National"],
    Party == "Labour Party" ~ parties_v["Labour"],
    Party == "New Zealand First Party" ~ parties_v["NZ First"],
    Party == "ACT New Zealand" ~ parties_v["ACT"],
    Party == "United Future" ~ parties_v["United Future"],
    Party == "Māori Party" ~ parties_v["Māori"]
  )) %>% 
  mutate(opa = 0.1 + 0.85 * ((PerCent - min(PerCent)) / (max(PerCent) - min(PerCent))))
electorate_winners$labeltext <-
  electorate_winners %>% 
  inner_join(nzhex) %>% 
  transpose() %>% 
  map(~ paste0(begin_label_text,
               begin_label_header, .x$electorate_name, end_label_header,
               .x$Party, " won with ", prettyNum(.x$Votes, big.mark=','), " votes.",
               "<br>",
               "<div style='width:95%'>",
               "<span style='color:", .x$party_colour, ";float:left'>", .x$PerCent, "%</span>",
               "<br clear='all' />",
               "<span style='background:", .x$party_colour, ";width:", .x$PerCent, "%;float:left'>&nbsp;</span>",
               "</div>",
               end_label_text) %>% 
        HTML())

leaflet(options = getLeafletOptions(5.8, 5.8)) %>% 
  addPolygons(
    data = inner_join(nzhex, electorate_winners),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Electorate Seats',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addLabelOnlyMarkers(
    data = nzhex %>% st_centroid(),
    label = ~abbr,
    labelOptions = labelOptions(
      noHide = 'T', textOnly = T,
      offset=c(-6,-11))
  ) %>% 
  onRender(
    "function(el, t) {
      var myMap = this;
      // get rid of the ugly grey background
      myMap._container.style['background'] = '#ffffff';
    }"
  )
```

***

Each hexagon is coloured for the party of the winning candidate. The strength of colour tracks
the percentage of votes the winner received --- from
`r electorate_winners %>% top_n(-1, PerCent) %$% PerCent`% for the
`r electorate_winners %>% top_n(-1, PerCent) %$% Party` in
`r electorate_winners %>% top_n(-1, PerCent) %>% inner_join(nzhex) %$% electorate_name`
to
`r electorate_winners %>% top_n(1, PerCent) %$% PerCent`% for the
`r electorate_winners %>% top_n(1, PerCent) %$% Party` in
`r electorate_winners %>% top_n(1, PerCent) %>% inner_join(nzhex) %$% electorate_name`.


### Party Votes: How are each party's voters distributed? Do they have broad or localised support?

```{r hex_parties}
party_votes <-
  GE2017 %>% 
  filter(VotingType == "Party") %>% 
  group_by(id, Party) %>% 
  summarise(Votes = sum(Votes)) %>% 
  ungroup() %>% 
  inner_join(party_voters_2017) %>% 
  mutate(PerCent = round(Votes / Voters * 100, 1)) %>% 
  mutate(party_colour = case_when(
    Party == "National Party" ~ parties_v["National"],
    Party == "Labour Party" ~ parties_v["Labour"],
    Party == "New Zealand First Party" ~ parties_v["NZ First"],
    Party == "ACT New Zealand" ~ parties_v["ACT"],
    Party == "United Future" ~ parties_v["United Future"],
    Party == "Māori Party" ~ parties_v["Māori"],
    Party == "Green Party" ~ parties_v["Green"],
    Party == "The Opportunities Party (TOP)" ~ parties_v["TOP"],
    Party == "Aotearoa Legalise Cannabis Party" ~ parties_v["Legalise Cannabis"],
    Party == "Ban1080" ~ parties_v["Ban 1080"],
    Party == "Conservative" ~ parties_v["Conservative"],
    Party == "Internet Party" ~ parties_v["Internet"],
    Party == "MANA" ~ parties_v["Mana"]
  )) %>% 
  group_by(Party) %>% 
  mutate(mean_votes = mean(Votes),
         opa = case_when(
           Votes < 0.8 * mean_votes ~ 0.15,
           Votes < 0.95 * mean_votes ~ 0.3,
           Votes <= 1.05 * mean_votes ~ 0.5,
           Votes <= 1.2 * mean_votes ~ 0.7,
           Votes > 1.2 * mean_votes ~ 0.85
         ),
         MaxPC = ceiling(max(PerCent))
  ) %>% 
  ungroup()

party_votes_2014 <-
  GE2014 %>% 
  filter(VotingType == "Party") %>% 
  group_by(id, Party) %>% 
  summarise(Votes2014 = sum(Votes)) %>% 
  ungroup() %>% 
  inner_join(party_voters_2014) %>% 
  mutate(PerCent2014 = round(Votes2014 / Voters2014 * 100, 1)) %>% 
  group_by(Party) %>% 
  mutate(MaxPC2014 = ceiling(max(PerCent2014))) %>% 
  ungroup()

party_votes$labeltext <-
  party_votes %>% 
  left_join(
    party_votes_2014 %>% 
      filter(Party != "Internet MANA") %>% 
      bind_rows(
        party_votes_2014 %>% 
          filter(Party == "Internet MANA") %>% 
          mutate(Party = "MANA")
      ) %>% 
      bind_rows(
        party_votes_2014 %>% 
          filter(Party == "Internet MANA") %>% 
          mutate(Party = "Internet Party")
      )
    ) %>% 
  group_by(Party) %>% 
  mutate(MaxPC = if_else(is.na(MaxPC2014), MaxPC, max(MaxPC, MaxPC2014))) %>% 
  ungroup() %>% 
  inner_join(nzhex) %>% 
  transpose() %>% 
  map(~ paste0(begin_label_text_200,
               begin_label_header, .x$electorate_name, end_label_header,
               prettyNum(.x$Votes, big.mark=','), " votes.",
               "<br>",
               "<div style='width:95%'>",
               "<span style='float:left'>2017: ", .x$PerCent, "%</span>",
               "<br clear='all' />",
               "<span style='background:", .x$party_colour, ";width:", round(.x$PerCent / .x$MaxPC * 100), "%;float:left'>&nbsp;</span>",
               if (is.na(.x$MaxPC2014)) {""} else {
                 paste0(
                   "<br clear='all' />",
                   "<span style='float:left'>2014: ", .x$PerCent2014, "%</span>",
                   "<br clear='all' />",
                   "<span style='background:", .x$party_colour, ";width:", round(.x$PerCent2014 / .x$MaxPC * 100), "%;float:left'>&nbsp;</span>"
                 )
               },
               "</div>",
               end_label_text) %>% 
        HTML())

leaflet(options = getLeafletOptions(5.8, 5.8)) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "National Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'National 44.4%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Labour Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Labour 36.9%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "New Zealand First Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'NZ First 7.2%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Green Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Green 6.3%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "The Opportunities Party (TOP)")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'TOP 2.4%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Māori Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Māori 1.2%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "ACT New Zealand")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'ACT 0.5%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addLabelOnlyMarkers(
    data = nzhex %>% st_centroid(),
    label = ~abbr,
    labelOptions = labelOptions(
      noHide = 'T', textOnly = T,
      offset=c(-6,-11))
  ) %>% 
  addLayersControl(
    baseGroups = c("National 44.4%", "Labour 36.9%", "NZ First 7.2%", "Green 6.3%",
                   "TOP 2.4%", "Māori 1.2%", "ACT 0.5%"),
    options = layersControlOptions(collapsed = FALSE)
  ) %>% 
  onRender(
    "function(el, t) {
      var myMap = this;
      // get rid of the ugly grey background
      myMap._container.style['background'] = '#ffffff';
    }"
  )
```

***

Choose a party and see the tiles shaded relative to the number of voters. There are
five shades: within 5% of the average, between 5-20% over (or under) average, and more than 20% over (or under) the average.
The mouseover compares the 2017 and 2014 percentages.

Note that because it's based on absolute numbers rather than percentages, the Māori
electorates may be unexpectedly lighter because of their lower turnouts.

National and Labour have fairly complementary patterns, as do NZ First and the Greens.



### Coalition votes: Which electorates were "won" by the incumbent government?

```{r hex_coalition}
coalition_votes <-
  party_votes %>% 
  filter(Party %in% c("National Party", "United Future", "Māori Party", "ACT New Zealand")) %>% 
  group_by(id) %>% 
  summarise(Votes = sum(Votes),
            Voters = min(Voters),
            party_colour = min(party_colour)) %>% # min=National
  mutate(PerCent = round(Votes / Voters * 100, 1),
         MaxPC = ceiling(max(PerCent))) %>% 
  ungroup() %>% 
  mutate(win = PerCent > 50)

coalition_votes_2014 <-
  party_votes_2014 %>% 
  filter(Party %in% c("National Party", "United Future", "Māori Party", "ACT New Zealand")) %>% 
  group_by(id) %>% 
  summarise(Votes2014 = sum(Votes2014),
            Voters2014 = min(Voters2014)) %>%
  mutate(PerCent2014 = round(Votes2014 / Voters2014 * 100, 1),
         MaxPC2014 = ceiling(max(PerCent2014))) %>% 
  ungroup() %>% 
  mutate(win2014 = PerCent2014 > 50)

coalition_votes$labeltext <-
  coalition_votes %>% 
  left_join(coalition_votes_2014) %>% 
  inner_join(nzhex) %>% 
  transpose() %>% 
  map(~ paste0(begin_label_text_200,
               begin_label_header, .x$electorate_name, end_label_header,
               prettyNum(.x$Votes, big.mark=','), " votes.",
               "<br>",
               "<div style='width:95%'>",
               "<span style='float:left'>2017: ", .x$PerCent, "%</span>",
               "<br clear='all' />",
               "<span style='background:", .x$party_colour, ";width:", round(.x$PerCent / .x$MaxPC * 100), "%;float:left'>&nbsp;</span>",
               if (is.na(.x$MaxPC2014)) {""} else {
                 paste0(
                   "<br clear='all' />",
                   "<span style='float:left'>2014: ", .x$PerCent2014, "%</span>",
                   "<br clear='all' />",
                   "<span style='background:", .x$party_colour, ";width:", round(.x$PerCent2014 / .x$MaxPC * 100), "%;float:left'>&nbsp;</span>"
                 )
               },
               "</div>",
               end_label_text) %>% 
        HTML())

leaflet(options = getLeafletOptions(5.8, 5.8)) %>% 
  addMinicharts(
    lng = nzhex %>% 
      arrange(id) %>% 
      st_centroid() %>% 
      st_coordinates() %>% 
      extract(, "X"),
    lat = nzhex %>% 
      arrange(id) %>% 
      st_centroid() %>% 
      st_coordinates() %>% 
      extract(, "Y"),
    chartdata = coalition_votes %>% 
      arrange(id) %>% 
      transmute(Coalition = PerCent, `"Change"` = 100 - PerCent),
    type = "pie",
    colorPalette = c("darkblue", "orange")
  ) %>%
  addPolygons(
    data = inner_join(nzhex, coalition_votes),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Coalition Voters',
    fillColor = "black", fillOpacity = ~if_else(win, 0.2, 0),
    highlightOptions = highlightOptions(weight = 4)
  ) %>%
  addLabelOnlyMarkers(
    data = nzhex %>% st_centroid(),
    label = ~abbr,
    labelOptions = labelOptions(
      style = list("color" = "white"),
      noHide = 'T', textOnly = T,
      offset=c(-6,-11))
  ) %>%
  onRender(
    "function(el, t) {
      var myMap = this;
      // get rid of the ugly grey background
      myMap._container.style['background'] = '#ffffff';
    }"
  )
```

***

The incumbant government is a four-party coalition between National (47.0% in 2014),
Māori (1.3%), ACT (0.7%) and United Future (0.2%). Some have argued that the election
was a referendum on the government and that the majority voted against it.

Though I'm no fan of pie charts in general, I think they're reasonable here to show the
proportion of votes for and against coalition parties, with the latter arguably in
favour of some sort of change.

The coalition received a majority in just 
`r sum(coalition_votes$win)` of 71 electorates, down from
`r sum(coalition_votes_2014$win2014)` in 2014.

They increased their vote in
`r inner_join(coalition_votes, coalition_votes_2014) %>% filter(PerCent > PerCent2014) %>% nrow()`
electorates, but none by more than 4% (the 2014 national vote for the collapsed
Conservative party).
The decreases ranged up to
`r inner_join(coalition_votes, coalition_votes_2014) %>% inner_join(nzhex) %>% arrange(PerCent2014 - PerCent, PerCent) %>% tail(1) %$% {PerCent2014 - PerCent}`%
(in
`r inner_join(coalition_votes, coalition_votes_2014) %>% inner_join(nzhex) %>% arrange(PerCent2014 - PerCent, PerCent) %>% tail(1) %$% electorate_name`).


### Tiny Party Votes: Are the sub 0.5% parties evenly ignored, or do they have localised support?

```{r hex_minnows}
leaflet(options = getLeafletOptions(5.8, 5.8)) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Aotearoa Legalise Cannabis Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Cannabis 0.3%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Conservative")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Conservative 0.2%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "MANA")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'MANA 0.1%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Ban1080")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Ban1080 0.1%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "New Zealand People's Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = "NZ People's 0.1%",
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "United Future")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'United Future 0.1%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "NZ Outdoors Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'NZ Outdoors 0.1%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Democrats for Social Credit")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Social Credit 0.0%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addPolygons(
    data = nzhex %>% 
      inner_join(party_votes %>% filter(Party == "Internet Party")),
    label = ~labeltext,
    weight=2, color='#000000', group = 'Internet 0.0%',
    fillOpacity = ~opa, opacity = 1, fillColor = ~party_colour,
    highlightOptions = highlightOptions(weight = 4)
  ) %>% 
  addLabelOnlyMarkers(
    data = nzhex %>% st_centroid(),
    label = ~abbr,
    labelOptions = labelOptions(
      noHide = 'T', textOnly = T,
      offset=c(-6,-11))
  ) %>% 
  addLayersControl(
    baseGroups = c("Cannabis 0.3%", "Conservative 0.2%", "MANA 0.1%", "Ban1080 0.1%", "NZ People's 0.1%",
                   "United Future 0.1%", "NZ Outdoors 0.1%", "Social Credit 0.0%", "Internet 0.0%"),
    options = layersControlOptions(collapsed = FALSE)
  ) %>% 
  onRender(
    "function(el, t) {
      var myMap = this;
      // get rid of the ugly grey background
      myMap._container.style['background'] = '#ffffff';
    }"
  )
```

***

More people cast an "informal" vote than voted for any of these parties.

MANA unsurprisingly have most of their support in the Māori seats.
I'm seriously wondering if they got some votes in the Mana electorate from people assuming they
had a local focus.

The rest of the distributions are pretty much what I expected, though when a party doesn't break
100 votes in any electorate (as is the case with the bottom four) it's easy to overinterpret the noise.

The biggest surprise to me was that every party got at least 1 vote in every electorate.





<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>