This Week’s Widget - mapshaperWidget

Even a geo-novice like me can quickly see the amazing power of mapshaper and its web gui and even use it immediately. The author has clearly focused on a speedy, intuitive, but still full-feautured user experience. To take it one step further though, let’s try to integrate mapshaper seamlessly into our R geospatial workflow.

For now, I chose the awkward name mapshaperWidget, since “mapshaperr” is a little hard and there already is rmapshaper to use the non-gui node.js version. I’m thinking if ok with the author, eventually I can fold mapshaperWidget into rmapshaper.

mapshaper is an entire website, so converting it to an htmlwidget is much different from most of the other conversions. As of now, mapshaperWidget is not composeable and will not integrate into rmarkdown, since it completely takes over the page. In later iterations, I hope to contain it.

mapshaperWidget relies on geojsonio. Thanks to the contributors for this very helpful package. I also enjoyed adding some lazy magic from Hadley Wickham’s lazyeval. Also, one last thanks to RStudio for integrating RStudio Viewer, which allows for a smooth all-in-one window experience.

Quick Installation

mapshaperWidget is not yet on CRAN, so for now please install with devtools::install_github.

devtools::install_github("timelyportfolio/mapshaper_htmlwidget")

Examples

Examples usually help more than text. Let’s see if this is true.

# devtools::install_github("timelyportfolio/mapshaperWidget")
library(mapshaperWidget)

alabama <- paste0(
    readLines(
        "https://raw.githubusercontent.com/glynnbird/usstatesgeojson/master/alabama.geojson"
    )
    ,collapse = " "
)
mapshaper(alabama)

# devtools::install_github("timelyportfolio/mapshaperWidget")
library(mapshaperWidget)
library(tmap)

data(World)
mapshaper(World)

# devtools::install_github("timelyportfolio/mapshaperWidget")
library(mapshaperWidget)
library(spdep)
library(maptools)

auckland <- readShapePoly(
  system.file(
    "etc/shapes/auckland.shp"
    ,package="spdep"
  )[1]
)
wheat <- readShapePoly(
  system.file(
    "etc/shapes/wheat.shp"
    ,package="spdep"
  )[1]
)
columbus <- readShapePoly(
  system.file(
    "etc/shapes/columbus.shp"
    ,package="spdep"
  )[1]
)

mapshaper( auckland, wheat, columbus )

Next…

This is certainly not all, but I’ll stop here. Next step will be bidirectional communication with Shiny and as said earlier some containment for composeability. As always, I would love to work together if you have some ideas.

Thanks

Thanks so much to Matthew Bloch from the New York Times for mapshaper. It is really good!

As always, thanks to

• Ramnath Vaidyanathan and RStudio for htmlwidgets
• all the contributors to R and JavaScript