Search code examples
rhighchartsheatmapr-highcharter

R highcharter: putting heatmaps into motion


I'm trying to use the highcharter R package "Motion Plugin", to make a Motion Chart for a heatmap. I.e. I would like a heatmap to change over time, using a slider with a play/pause button (see links below).

I'm able to create a simple heatmap, for a specific year, e.g.:

df <- tibble(year = c(rep(2016, 6), rep(2017, 6)),
         xVar = rep(c("a", "a", "b", "b", "c", "c"), 2),
         yVar = rep(c("d", "e"), 6),
         heatVar = rnorm(12))

df %>%
  filter(year == 2016) %>%
  hchart(type = "heatmap", hcaes(x = xVar, y = yVar, value = heatVar)) %>%
  hc_legend(layout = "vertical", verticalAlign = "top", align = "right")

highcharter heatmap

However, I'm struggling with making this a Motion Chart (sliding across 2016, 2017 in this example), using the hc_motion(enabled = TRUE, ...) function.

I have read and followed these links:

https://www.r-bloggers.com/adding-motion-to-choropleths/

http://jkunst.com/highcharter/plugins.html

But no matter how I define my series, I do not get the expected result. Anyone can point me how the xVar, yVar series should be defined and the hc_motion function be used to make it work?


UPDATE:

Following this answer I managed to this using shiny, but I'd still prefer to avoid this solution:

server <- shinyServer(function(input, output) {

  output$heatmap <- renderHighchart({

  df <- tibble(year = c(rep(2016, 6), rep(2017, 6)),
             xVar = rep(c("a", "a", "b", "b", "c", "c"), 2),
             yVar = rep(c("d", "e"), 6),
             heatVar = rnorm(12))

  # filter data based on selected year
  df.select <- dplyr::filter(df, year == input$year) 

  # chart
  hchart(df.select, type = "heatmap", hcaes(x = xVar, y = yVar, value = heatVar))



  })

})

ui <- shinyUI(fluidPage(

  # Application title
  titlePanel("Highcharts Heatmap Motion Chart"),

  # Sidebar with a slider input for the selected year
  sidebarLayout(
    sidebarPanel(
      sliderInput("year",
                  "Year:",
                  min = 2016,
                  max = 2017,
                  step = 1,
                  value = 2016,
                  animate = TRUE,
                  sep = "")
    ),

    # Show a bubble plot for the selected year
    mainPanel(
      highchartOutput("heatmap")
    )
  )
))

shinyApp(ui = ui, server = server)

Solution

  • The approach surely is not the cleanest, because it's needed to create the initial position (like a standard chart) then create te sequences for each point.

    http://rpubs.com/jbkunst/questions-42945062

    So the structure to add the motion plugin would be:

    Simulate data


    library(highcharter)
    library(dplyr)
    library(purrr)
    
    
    years <- 10
    nx <- 5
    ny <- 6
    df <- data_frame(year = rep(c(2016 + 1:years - 1), each = nx * ny), xVar = rep(1:nx, 
      times = years * ny), yVar = rep(1:ny, times = years * nx))
    
    df <- df %>% group_by(xVar, yVar) %>% mutate(heatVar = cumsum(rnorm(length(year))))
    

    Get initial values

    df_start <- df %>% arrange(year) %>% distinct(xVar, yVar, .keep_all = TRUE)
    df_start
    #> Source: local data frame [30 x 4]
    #> Groups: xVar, yVar [30]
    #> 
    #>     year  xVar  yVar    heatVar
    #>    <dbl> <int> <int>      <dbl>
    #> 1   2016     1     1  0.5894443
    #> 2   2016     2     2 -1.0991727
    #> 3   2016     3     3  1.1209292
    #> 4   2016     4     4  0.4936719
    #> 5   2016     5     5 -0.4614157
    #> # ... with 25 more rows
    

    Grouping for the fixed variables to create a list with the sequence

    df_seqc <- df %>% group_by(xVar, yVar) %>% do(sequence = list_parse(select(., 
      value = heatVar)))
    df_seqc
    #> Source: local data frame [30 x 3]
    #> Groups: <by row>
    #> 
    #> # A tibble: 30 × 3
    #>     xVar  yVar    sequence
    #> *  <int> <int>      <list>
    #> 1      1     1 <list [10]>
    #> 2      1     2 <list [10]>
    #> 3      1     3 <list [10]>
    #> 4      1     4 <list [10]>
    #> 5      1     5 <list [10]>
    #> # ... with 25 more rows
    

    Join

    data <- left_join(df_start, df_seqc)
    #> Joining, by = c("xVar", "yVar")
    data
    #> Source: local data frame [30 x 5]
    #> Groups: xVar, yVar [?]
    #> 
    #>     year  xVar  yVar    heatVar    sequence
    #>    <dbl> <int> <int>      <dbl>      <list>
    #> 1   2016     1     1  0.5894443 <list [10]>
    #> 2   2016     2     2 -1.0991727 <list [10]>
    #> 3   2016     3     3  1.1209292 <list [10]>
    #> 4   2016     4     4  0.4936719 <list [10]>
    #> 5   2016     5     5 -0.4614157 <list [10]>
    #> # ... with 25 more rows
    

    And chart

    limits <- (unlist(data$sequence)) %>% {
      c(min(.), max(.))
    }
    limits
    #> [1] -5.332709  6.270384
    
    hc1 <- hchart(data, type = "heatmap", hcaes(x = xVar, y = yVar, value = heatVar))
    
    hc2 <- hchart(data, type = "heatmap", hcaes(x = xVar, y = yVar, value = heatVar)) %>% 
      hc_motion(enabled = TRUE, series = 0, startIndex = 0, labels = unique(df$year)) %>% 
      hc_legend(layout = "vertical", verticalAlign = "top", align = "right") %>% 
      hc_colorAxis(min = limits[1], max = limits[2])