/sites_viz

Primary LanguageR

sites_viz

Pacakges

The following packages are used for this project:

* shiny
* shinydashboard
* leaflet
* dplyr
* plyr
* sf
* sp
* leafgl
* DT

Note

The package leafgl is used in conjunction with leaflet to create maps. It delivers fantastic performance.

While leaflet is great, it does need help when it comes to performance, especially when the number of data points that we have is astronomical. The reason why leaflet suffers from having many data points is that each point is an individual DOM element. Browsers simply cannot cope well with an HTML page that's got thousands of DOM elements.

The downside with leafgl is that it is pre-alpha and under heavy development at the moment. As such, I recommend download a working version as early as possible. Here is a link to leafgl's git repo: https://github.com/r-spatial/leafgl

In the case that leafgl can no longer function, I would sugget future developers to look into adding each point (marker) onto a canvas first and then render the canvas. This avoids having tens of thousands of DOM elements.

Structure

ui.R
server.R
data/
src/

For the time being, all the datasets are provided in a .csv format and are stored within the src/ folder. Every dataset has the following three columns:

  • X: index of a data point, which we use as ID for each data point
  • o_longitude_GDA94
  • o_latitude_GDA94

In the src/ folder, R scripts that ui.R and server.R depend on can be found.

To run the app, open either ui.R or server.R in RStudio. After the file has been opened, you will find a special Run App button on the top right corner of the script. Clicking on Run App will start up the application.

Alternatively, one could use shift + command + return on Mac or shift + control + enter on Windows.

To run the app in a docker container:

  1. Build the docker image: docker build -t my-shiny-app .
  2. And then run a container using that image: docker run --rm -p 3838:3838 my-shiny-app
  3. The shiny app would then be accessible on the host machine at the URL: http://localhost:3838/

For more details on the project, please refer to the design document.