Problem with ggplot geom-tile convert to plotly chartusing ggplotly

Question

I tried to convert the ggplot geom-tile chart to plotly by using ggplotly function. However, I realized that the outcomes are different. Please refer for the link below to view the difference. Besides that, ggplotly chart also missing colorbar. Please help.

Image: Difference between ggplot and ggplotly chart

This is the code that I referred to : https://www.r-graph-gallery.com/285-waffer-map.html

Code:

madeUp=read.table("https://raw.githubusercontent.com/holtzy/R-graph-gallery/master/DATA/madeUp.csv", sep=",", header=T)


library(tidyverse)


theData <- madeUp %>% 
  group_by(X.Axis, Y.Axis, Group) %>% 
  dplyr::summarize(statistic=mean(randVals, na.rm = TRUE))

fig <- ggplot(theData, aes(X.Axis, Y.Axis)) +

  coord_cartesian(xlim = c(0,20), ylim = c(0,20)) +
  scale_x_continuous(breaks = seq(0,20)) +
  scale_y_continuous(breaks = seq(0,20))+

  geom_tile(aes(fill=statistic))+
  guides(fill=guide_legend(title='Legend'))+

  theme(
    panel.background = element_rect(fill= 'white', color = 'white'),
    panel.grid.major = element_line(color='#E0E0E0'),
    panel.grid.minor = element_line(color='#E0E0E0')
  )+

  ggtitle('Wafer Map')+
  facet_wrap(~Group)+
  scale_fill_gradientn(colors = rainbow(100))

#Convert to ggplotly chart
fig <- ggplotly(fig)

Thanks

Answer 1

Looks like you stumbled over a bug (or two) in ggplotly (maybe you should raise an issue on github).

1. The first issue is that the "gaps" in the dataset get lost when converting the ggplot via ggplotly.

2. The second issue is that ggplotly fails to convert the binned colorbar added by guides(fill=guide_legend(title='Legend')).

As a workaround for

1. the first issue you could expand the dataset to include all combinations of X.Axis, Y.Axis and Group.
2. the second issue you could remove the binned colorbar and replace it by a continuous colorscale.

Not perfect but this way the conversion via ggplotly gives you the correct plot and a legend. Try this:

madeUp=read.table("https://raw.githubusercontent.com/holtzy/R-graph-gallery/master/DATA/madeUp.csv", sep=",", header=T)

theData <- madeUp %>% 
  group_by(X.Axis, Y.Axis, Group) %>% 
  dplyr::summarize(statistic=mean(randVals, na.rm = TRUE)) %>% 
  ungroup()

# Expand the Dataset to includ all Combinations of X.Axis, Y.Axis and Group
theData1 <- tidyr::expand_grid(X.Axis = 0:49, Y.Axis = 0:30, Group = LETTERS[1:6]) %>% 
  dplyr::left_join(theData)

fig <- ggplot(theData1, aes(X.Axis, Y.Axis)) +
  coord_cartesian(xlim = c(0,20), ylim = c(0,20)) +
  scale_x_continuous(breaks = seq(0,20)) +
  scale_y_continuous(breaks = seq(0,20))+
  geom_tile(aes(fill=statistic))+
  # Remove the binned colorbar
  # guides(fill=guide_legend(title='Legend'))+
  labs(fill = "Legend") +
  theme(
    panel.background = element_rect(fill= 'white', color = 'white'),
    panel.grid.major = element_line(color='#E0E0E0'),
    panel.grid.minor = element_line(color='#E0E0E0')
  )+

  ggtitle('Wafer Map')+
  facet_wrap(~Group)+
  # in case of ggplot2: Set the fill color for NA to "transparent"
  scale_fill_gradientn(colors = rainbow(100), na.value = "transparent")
fig

ggplotly(fig)