dot heatmap

[antoine4ucsd]

Hello thank you for your impressive set of tools. I am trying to make a dot heatmap like below (in ggplot2) I tried a scatter plot but does not seem to work with discete y axis? Is there a workaround? thank you!

[rdatasculptor]

Maybe you can put your attempt, the code here? It helps finding a solution

[antoine4ucsd]

thank you. here;s an example.

# scatter not working (ideally woule like the size ~ Avg and color ~ Pct but fine if only ~ Pct
 data|>
        group_by(pid)|>
        e_charts(Gene,timeline = TRUE) |>
        e_scatter(predicted.celltype,Avg) |>
        # e_theme("infographic") |>
        e_legend(show=T)|>
        e_x_axis(axisLabel = list(rotate = 90))
# heatmap working 
data|>
        group_by(pid)|>
        e_charts(Gene,timeline = TRUE) |>
        e_heatmap(predicted.celltype, Avg) |> 
        e_theme("infographic") |>  
        e_visual_map(Avg)|>
[data.csv](https://github.com/JohnCoene/echarts4r/files/13260061/data.csv)

        e_legend(show=T)|>
        e_x_axis(axisLabel = list(rotate = 90))

thank you!

[rdatasculptor]

Something like this?

data <- read.csv("https://github.com/JohnCoene/echarts4r/files/13260061/data.csv") %>%
  mutate(predicted.celltypeNumber = as.numeric(factor(predicted.celltype )) )

data|>
  group_by(pid)|>
  e_charts(Gene, timeline = TRUE) |>
  e_scatter(predicted.celltypeNumber, Avg, bind = predicted.celltype,
            itemStyle = list(color = "darkgreen")) |>
  e_visual_map(predicted.celltypeNumber, dimension = 1,
               inRange = list(color = c("darkgreen", "lightgreen")),
               inverse = TRUE) |>
  e_legend(show=T)|>
  e_x_axis(show = FALSE, axisLabel = list(rotate = 90)) |>
  e_y_axis(show = FALSE)

[antoine4ucsd]

Promising I will play around this code and get back to you Thank you so much !

[antoine4ucsd]

I looked into the solution proposed by helgasoft with echarty and it does the job very nicely. thank you all for your contributions. this is exciting to see such a great community.

[antoine4ucsd]

when trying your code, I noticed the y axis were not laeblined. Ideally I would need to discrete (text) label, to be shown.

I assume it starts with something around those lines?

data|>
        group_by(pid)|>
        e_charts(Gene, timeline = TRUE) |>
        e_scatter(predicted.celltypeNumber, Avg, bind = predicted.celltype,
                  itemStyle = list(color = "darkgreen")) |>
        e_visual_map(predicted.celltypeNumber, dimension = 1,
                     inRange = list(color = c("darkgreen", "lightgreen")),
                     inverse = TRUE) |>
        e_legend(show=T)|>
        e_x_axis(show = T, axisLabel = list(rotate = 90)) |>
        e_y_axis(show = T,
                 formatter = htmlwidgets::JS("
            function(params){
              return('<strong>' + params.name )
                      }
          "))

I wonder if you could help or guide me to where I can learn these 'advanced' options also is there a way to scale the dot size limits? can the size and color be scaled differently? one being the Pct and the other being the Avg?

I have other question but will stop here for now ;-)

thank you!

[rdatasculptor]

Your first example did not contain labels. I tried the chart look like that one. But ofcourse I understand what you would like to see. By the way, I could not find the echarty solution you mentioned.

[antoine4ucsd]

sorry about that. the label would have been predicted.celltype (currenty, the label is the predicted.celltype.number). does that help?

echarty solution is here: https://gist.github.com/helgasoft/e914fe6f4a9ed9407f78e41cb18b3aaa#gistcomment-4750586

[rdatasculptor]

echarty is highly customizable. However at least a part of what is done there, can be easlily done by echarts4r as well::

library(dplyr)
library(echarts4r)
data <- read.csv("https://github.com/JohnCoene/echarts4r/files/13260061/data.csv") %>%
  mutate(Avg = ifelse(Avg == 0, NA, Avg))

data|>
  group_by(pid)|>
  e_charts(Gene, timeline = TRUE) |>
  e_heatmap(predicted.celltype, Avg) |>
  e_add_unnested("symbolSize", Avg) |>
  e_visual_map(Avg) |>
  e_legend(show=T) |>
  e_x_axis(axisLabel = list(rotate= 90))

[antoine4ucsd]

yes, for the heatmap this is great. was just trying the dot plot with size ~Avg and color ~Pct. not sure if this is too complicated. anyway,thank you!

[rdatasculptor]

Sorry for misunderstanding your question(s).

Is this what you are looking for? You can play with the scale to give the bubles te size you prefer:

data <- read.csv("https://github.com/JohnCoene/echarts4r/files/13260061/data.csv") %>%
  mutate(Avg = ifelse(Avg == 0, NA, Avg))

data|>
  group_by(pid)|>
  e_charts(Gene, timeline = TRUE) |>
  e_scatter(predicted.celltype, Avg, scale = function(x) scales::rescale(x, to = c(10, 70))) |>
  e_visual_map(Avg, dimension = 2) |>
  e_x_axis(type = "category", axisLabel = list(rotate = 90)) |>
  e_y_axis(type = "category") |>
  e_data(data, predicted.celltypeNumber)

[antoine4ucsd]

I think this is it! I will play around it and confirm asap thank you!

[antoine4ucsd]

I confirm it works perfectly. Exactly what I was looking for. I need to get more familiar with some of the arguments above but this is great! thank you again

[rdatasculptor]

For me the excitement about echarts4r really grew once I realized the possibilities in building and tweaking charts just doesn't stop with the great job that @JohnCoene did with this package. You can do almost anything when getting familiar with https://echarts.apache.org/en/option.html#title itself. Almost any option or argument is reachable from within @JohnCoene's echarts4r functions. Powerful!

[antoine4ucsd]

cannot agree more. sky is the limit (and my skillset...) thank you all