How to: Weather Radials

Posted on March 23, 2016 by Joshua Kunst in R bloggers | 0 Comments

[This article was first published on Jkunst - R category, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

TLDR: Creating weather radials with highcharter and ggplot2.

I was surfing by the deep seas of the web and I found the Brice Pierre de la Briere’s blocks and I saw the weather radials which originally are a poster collection. Brice uses D3 and he used D3 very well and I love D3 but I’m in a rookie level to do something like him. D3 is not for everybody and surely not for me, I would love to lear more but family, work and R has priority over D3 so how can I do something like that. Well… We have R & highcarter. So let’s try.

We’ll use the same data as Brice [https://www.wunderground.com/].

df <- read_csv("http://bl.ocks.org/bricedev/raw/458a01917183d98dff3c/sf.csv")

df[1:4, 1:4]

date	Max TemperatureC	Mean TemperatureC	Min TemperatureC
2014-01-01	13	9	5
2014-01-02	17	12	6
2014-01-03	18	12	7
2014-01-04	19	13	6

names(df) <- names(df) %>% 
  str_to_lower() %>% 
  str_replace("\s+", "_")

df <- df %>% 
  mutate(id = seq(nrow(df)),
         date2 = as.Date(ymd(date)),
         tmstmp = datetime_to_timestamp(date2),
         month = month(ymd(date)))

dsmax <- df %>%
  select(x = tmstmp,
         y = max_temperaturec) %>% 
  list.parse3()
 
dsmin <- df %>% 
  select(x = tmstmp, y = min_temperaturec) %>% 
  list.parse3()

First try

Here we test and chart the data in the most simple way. A line time.

hc <- highchart() %>% 
  hc_chart(
    type = "line"
    ) %>%
  hc_xAxis(
    type = "datetime",
    tickInterval = 30 * 24 * 3600 * 1000,
    labels = list(format = "{value: %b}")
  ) %>% 
  hc_yAxis(
    min = 0,
    labels = list(format = "{value} C")
  ) %>% 
  hc_add_series(
    data = dsmax, name = "max"
  ) %>% 
  hc_add_series(
    data = dsmin, name = "min"
    ) %>% 
  hc_add_theme(
    hc_theme_smpl()
    )

hc

open

Everything seems fine.

Second Step

We’ change the type to column, stack and see what is the result

hc <- hc %>% 
  hc_chart(
    type = "column"
    ) %>% 
  hc_plotOptions(
    series = list(
      stacking = "normal"
    )
  )

hc

open

Not so close.

Final Step

If you see the previous chart we stacked so we sum the min and max and the data don’t reflect the value (min,max) what we want. So we need to create the difference between the max and min, and plot them with the min value and hiding using a transparent color.

And set polar = TRUE obviously.

dsmax <- df %>% 
  mutate(color = colorize_vector(mean_temperaturec, "A"),
         y = max_temperaturec - min_temperaturec) %>% 
  select(x = tmstmp,
         y,
         name = date,
         color,
         mean = mean_temperaturec,
         max = max_temperaturec,
         min = min_temperaturec) %>% 
  list.parse3()


# Some tooltips to make it a little *intercative*
x <- c("Min", "Mean", "Max")
y <- sprintf("{point.%s}", tolower(x))
tltip <- tooltip_table(x, y)

hc <- highchart() %>% 
  hc_chart(
    type = "column",
    polar = TRUE
  ) %>%
  hc_plotOptions(
    series = list(
      stacking = "normal",
      showInLegend = FALSE
    )
  ) %>% 
  hc_xAxis(
    gridLineWidth = 0.5,
    type = "datetime",
    tickInterval = 30 * 24 * 3600 * 1000,
    labels = list(format = "{value: %b}")
  ) %>% 
  hc_yAxis(
    max = 30,
    min = -10,
    labels = list(format = "{value} C"),
    showFirstLabel = FALSE
    ) %>% 
  hc_add_series(
    data = dsmax
  ) %>% 
  hc_add_series(
    data = dsmin,
    color = "transparent",
    enableMouseTracking = FALSE
  ) %>% 
  hc_add_theme(
    hc_theme_smpl()
  ) %>% 
  hc_tooltip(
    useHTML = TRUE,
    headerFormat = as.character(tags$small("{point.x:%d %B, %Y}")),
    pointFormat = tltip
  )

hc

open

Yay :D! A beautiful chart same as the d3 version and only using R. So sweet!

I’m happy with the result. This is not a standar chart but is a king of artistic. What do you think? Any other examples to test this type of chart?

Bonus Track: ggplot2 version

It’s really really easy to do this type of chart in ggplot2 using geom_linerange and geom_polar:

library("ggplot2")
library("viridis")
library("scales")

ggplot(df, aes(date2,
               ymin = min_temperaturec,
               ymax = max_temperaturec,
               color = mean_temperaturec)) + 
  geom_linerange(size = 1.3, alpha = 0.75) +
  scale_color_viridis(NULL, option = "A") +
  scale_x_date(labels = date_format("%b"), breaks = date_breaks("month")) + 
  ylim(-10, 35) + 
  labs(title = "San Francisco Wather Radial",
       subtitle = "It would be nice if someone do this with the animation package",
       caption = "Other example for ggplot2 vs base #boring but #fun",
       x = NULL, y = NULL) +
  coord_polar() + 
  theme_jbk() +
  theme(legend.position = "bottom")