How to: Weather Radials

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 & highcharter. So let’s try.

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

library(tidyverse)
library(highcharter)
library(lubridate)

df <- read_csv("https://gist.githubusercontent.com/bricedev/458a01917183d98dff3c/raw/f403fd2607be03b2d36fa024f1c6042730224927/sf.csv")

glimpse(df)

Rows: 365
Columns: 23
$ date                         <chr> "2014-1-1", "2014-1-2", "2014-1-3", "2014…
$ `Max TemperatureC`           <dbl> 13, 17, 18, 19, 19, 16, 16, 15, 16, 16, 1…
$ `Mean TemperatureC`          <dbl> 9, 12, 12, 13, 13, 11, 12, 13, 13, 12, 11…
$ `Min TemperatureC`           <dbl> 5, 6, 7, 6, 7, 7, 8, 11, 11, 9, 8, 8, 6, …
$ `Dew PointC`                 <dbl> 7, 7, 8, 8, 3, 7, 11, 11, 10, 9, 11, 8, 7…
$ `MeanDew PointC`             <dbl> 4, 4, 5, 6, 0, 4, 7, 9, 8, 8, 8, 6, 3, 3,…
$ `Min DewpointC`              <dbl> 2, 1, 1, 1, -3, 0, 4, 8, 6, 6, 6, 2, 0, -…
$ `Max Humidity`               <dbl> 83, 80, 80, 93, 65, 71, 86, 86, 89, 89, 9…
$ `Mean Humidity`              <dbl> 69, 58, 59, 65, 45, 61, 70, 79, 73, 75, 8…
$ `Min Humidity`               <dbl> 55, 36, 37, 36, 24, 51, 53, 72, 57, 60, 7…
$ `Max Sea Level PressurehPa`  <dbl> 1023, 1022, 1019, 1019, 1023, 1025, 1023,…
$ `Mean Sea Level PressurehPa` <dbl> 1022, 1020, 1016, 1016, 1021, 1023, 1022,…
$ `Min Sea Level PressurehPa`  <dbl> 1020, 1018, 1014, 1015, 1019, 1022, 1021,…
$ `Max VisibilityKm`           <dbl> 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 1…
$ `Mean VisibilityKm`          <dbl> 15, 16, 16, 9, 16, 15, 14, 15, 15, 15, 16…
$ `Min VisibilitykM`           <dbl> 11, 16, 14, 0, 16, 11, 11, 5, 8, 10, 10, …
$ `Max Wind SpeedKm/h`         <dbl> 14, 10, 23, 24, 23, 14, 24, 32, 40, 24, 2…
$ `Mean Wind SpeedKm/h`        <dbl> 1, 2, 6, 4, 8, 4, 7, 15, 19, 9, 10, 11, 4…
$ `Max Gust SpeedKm/h`         <dbl> 16, 10, 26, 29, 24, 16, 27, 39, 50, 27, 3…
$ Precipitationmm              <chr> "0.00", "0.00", "0.00", "0.00", "0.00", "…
$ CloudCover                   <dbl> 3, 3, 3, 2, 0, 5, 5, 7, 5, 4, 4, 2, 1, 0,…
$ Events                       <chr> NA, NA, NA, "Fog", NA, NA, "Rain", NA, NA…
$ WindDirDegrees               <dbl> 290, 146, 297, 266, 125, 75, 304, 281, 28…

df <- df %>%
  rename_all(str_to_lower) %>% 
  rename_all(str_replace, "\\s+", "_") %>% 
  mutate(
    id = seq(nrow(df)),
    date2 = as.Date(ymd(date)),
    tmstmp = datetime_to_timestamp(date2),
    month = month(ymd(date))
  )

dsmax <- df %>%
  select(tmstmp, max_temperaturec)
 
dsmin <- df %>% 
  select(tmstmp, min_temperaturec)

First try

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

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(dsmax,
                type = "line",
                hcaes(tmstmp, max_temperaturec),
                name = "max") %>%
  hc_add_series(dsmin,
                type = "line",
                hcaes(tmstmp, min_temperaturec),
                name = "min")

Everything seems fine. We now a little bit about the data

To polar coordinates and columnrange!

Now, we need to change to polar cordinates to get the “radial” part, this is get using the polar = TRUE argument in the hc_chart function. Then add the data in the columnrange type.

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

hchart(
  df,
  type = "columnrange",
  hcaes(x = date2, low = min_temperaturec, high = max_temperaturec, color = mean_temperaturec),
  showInLegend = FALSE
  ) %>% 
  hc_chart(
    polar = TRUE
  ) %>%  
  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} ºC"),
    showFirstLabel = FALSE
    ) %>% 
  hc_tooltip(
    useHTML = TRUE,
    headerFormat = as.character(tags$small("{point.x:%d %B, %Y}")),
    pointFormat = tltip
  )

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(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_c(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(legend.position = "bottom")

Nice!

Searching I found someone do this:

Always exist someone who did what you did before you.