charts_piecharts.Rmd

---
title: "Charts: Piecharts"
output:
  html_document:
    toc: true
    toc_float:
      collapsed: false
    includes:
      before_body: [includes/include_header.html, includes/include_header_navpage.html]
editor_options: 
  chunk_output_type: console
---
# What is a piechart?

Piecharts are useful visualisations for breaking down observations into a number of categories, the totals of which sum to 100% of the data. It is worth noting that it is often advisable to swap a barchart for a piechart, see 

- Categories: For instance, species of animal e.g. human or iguana
- Values: For instance, number of animals of each species

Below is a comparison between a piechart and barchart build using `R` and the `highcharter` library.

<!--html_preserve-->
<div class="row">
   <div class="col-md-6">

```{r basic_piechart, echo=FALSE, message=FALSE, warning=FALSE}
library("highcharter")
library("tidyverse")
my_data <- tribble(
  ~category, ~value,
  "human", 3,
  "non-human", 4,
  "snake", 6,
  "iguana", 8
)
my_data %>%
  hchart(
    type = "pie",
    hcaes(
      x = category,
      y = value
    )
  ) %>%
  hc_plotOptions(pie = list(dataLabels = list(enabled = FALSE))) %>%
  hc_size(width = "100%", height = "250px")
```

</div>

<div class="col-md-6">

```{r basic_barchart, echo=FALSE, message=FALSE, warning=FALSE}
library("highcharter")
library("tidyverse")
library("forcats")
my_data <- tribble(
  ~category, ~value,
  "human", 3,
  "non-human", 4,
  "snake", 6,
  "iguana", 8
)
my_data %>%
  arrange(desc(value)) %>%
  mutate(category = fct_reorder(category, value)) %>%
  hchart(
    type = "bar",
    hcaes(
      x = fct_reorder(category, rev(value)),
      y = value
    )
  ) %>%
  hc_plotOptions(pie = list(dataLabels = list(enabled = FALSE))) %>%
  hc_xAxis(title = list(text = "Category")) %>%
  hc_size(width = "100%", height = "250px")
```

</div>

</div>
<!--/html_preserve-->

# Limitations of piecharts

Research in graphical perception - the visual decoding of information encoded in graphs - suggests that piecharts are often an inefficient option for visualising the differences between many categories; as it is difficult for the visual perception system to judge small differences in the sizes of sections in a piechart. It's also difficult to get exact measurements from piecharts, in the example above the barchart makes it clear that half as many observations are "non-human" as are categorised as "iguana".

Often you will simply hear the advice "never use a piechart, instead use a barchart" based on the work of [Cleveland and McGill](http://doi.org/10.2307/2288400){target='_blank'} (amongst others). But blanket bans on piecharts are definitely overkill. There are many examples of where piecharts can be useful and a lot of variation possible in the design of piecharts, here are some useful resources and below there's a very simple comparison between a piechart and donut chart.

- EagerEyes.org blogpost on [designing effective piecharts](https://eagereyes.org/techniques/pie-charts)
- Study into [judging error in piechart visualisations](https://kosara.net/papers/2016/Kosara-EuroVis-2016.pdf)
- Study into [differences between pie, area and sector charts](https://kosara.net/papers/2016/Skau-EuroVis-2016.pdf)

<!--html_preserve-->
<div class="row">
   <div class="col-md-6">

```{r echo=FALSE, message=FALSE, warning=FALSE, paged.print=FALSE}
library("highcharter")
library("tidyverse")
my_data <- tribble(
  ~category, ~value,
  "human", 3,
  "non-human", 4,
  "snake", 6,
  "iguana", 8
)
my_data %>%
  hchart(
    type = "pie",
    hcaes(
      x = category,
      y = value
    )
  ) %>%
  hc_plotOptions(pie = list(dataLabels = list(enabled = TRUE))) %>%
  hc_size(width = "350px", height = "250px") %>%
  hc_plotOptions(pie = list(
            innerSize =  100,
            depth =  45
        ), series = list(animation = FALSE))
```
</div>
   <div class="col-md-6">
```{r echo=FALSE, message=FALSE, warning=FALSE, paged.print=FALSE}
library("highcharter")
library("tidyverse")
my_data <- tribble(
  ~category, ~value,
  "human", 3,
  "non-human", 4,
  "snake", 6,
  "iguana", 8
)
my_data %>%
  hchart(
    type = "pie",
    hcaes(
      x = category,
      y = value
    )
  ) %>%
  hc_plotOptions(pie = list(dataLabels = list(enabled = TRUE))) %>%
  hc_size(width = "350px", height = "250px") %>%
  hc_plotOptions(series = list(animation = FALSE))
```

</div>
</div>

<!--/html_preserve-->