Improving Performance of a leaflet

Question

I am trying to plot a map using Leaflet on Flexdashboard, output : html. I have a 50,000 data points i needed to plot. i tried using addCircleMarker(lng,lat) and it works in Base R but when the output is generated in an HTML file the performance of the map is so slow. I also tried using leafletOptions(preferCanvas = TRUE)but the performance is still very slow.

I also tried using a different library(leaflet.glify) to improve performance as leaflet.glifycan handle large dataset and used addGlifyPoints(data, color = cbind(0, 0, 0.1), group = "All cases") to plot the map. but i could not run it in an HTML output. It provided me with below error.

pandoc: Could not determine mime type for `/Library/Frameworks/R.framework/Versions/3.5/Resources/library/leaflet.glify/htmlwidgets/Leaflet.glify/src/shader/fragment/dot.glsl'
CallStack (from HasCallStack):
  error, called at src/Text/Pandoc/SelfContained.hs:156:35 in pandoc-1.19.2.1-JIeRA5EnQv74mk86CvVbbp:Text.Pandoc.SelfContained
Error: pandoc document conversion failed with error 1
Execution halted

Any help please.

Answer 1

For leaflet.glify you need to set self_contained: false in your yaml header:

---
title: "testing gl"
output: 
  flexdashboard::flex_dashboard:
    vertical_layout: fill
    self_contained: false
---

### Chart 1

```{r}
library(mapview)
library(leaflet)
library(leaflet.glify)
library(sf)

n = 1e5

df1 = data.frame(id = 1:n,
                 x = rnorm(n, 10, 3),
                 y = rnorm(n, 49, 1.8))

pts = st_as_sf(df1, coords = c("x", "y"), crs = 4326)

options(viewer = NULL) # view in browser

system.time({
  m = leaflet() %>%
    addProviderTiles(provider = providers$CartoDB.DarkMatter) %>%
    addGlifyPoints(data = pts, group = "pts") %>%
    addMouseCoordinates() %>%
    setView(lng = 10.5, lat = 49.5, zoom = 6) %>% 
    addLayersControl(overlayGroups = "pts")
})

m
```

### Chart 2

```{r}
plot(cars)
```