Skip to content

Latest commit

 

History

History
70 lines (51 loc) · 2.93 KB

README.md

File metadata and controls

70 lines (51 loc) · 2.93 KB
Raw

Community Snow Observations Web App

Community Snow Observations's web application is an interactive platform for viewing and exploring data collected by the project.

Getting Started

CSO's web app is built using Facebook's React library, and Next.js, a framework for React development. Application state is managed using Redux, and Saga is utilzied to handle asynchronous updates.

Running in development

The following steps can be followed using either yarn or npm

  1. Install necessary packages
cd src
yarn install
  1. Create .env file
touch .env
  1. Add mapbox variables to env file
MAPBOX_TOKEN={INSERT_MAPBOX_TOKEN_HERE}
MAPBOX_STYLE={INSERT_MAPBOX_STYLE_URL_HERE}
  1. Run frontend server
yarn run dev
  1. Go to localhost:3000

Packaging as static site

  1. Build production version of site
yarn run build
  1. Export site
yarn run export
  1. Push site to S3 bucket (Requires AWS sign in). Modify the build script in package.json to use your bucket of choice, then run yarn run push

Configuration

Most configuration options, including the theme, and filters, are defined in the config folder of the application. Some hardcoded configuration still exists throughout the application, but the aim is to move this code into the config folder to make modifying or retheming the application. When complete, this would allow easy reconfiguration, and possible re-use by similar projects.

Filtering

Filters are used in the application to select relevant observations. Current filter types include range filters, geometry filters, and discrete (categorical) filters. In order to speed up filtering, complete recaulations are not performed at every filter update event. Rather, cached intermediate results are used to speed up performance, and many updates are rendered efficiently on the map using GPU processing, while CPU updates, which are used to update the analytics panel, are debounced for better performance. To aid this process, the filtering process is expressed as a computational graph, and updates are only calculated as needed, when application updates invalidate existing results.

UI

UI components are separated into three scales. Small, resuable components are defined within the components/atoms folders. Compound components are defined in components/molecules. Very large components are defined in components/cells. This approach roughly conforms to the Atomic Design Methodology.

TODO

  • CSV downloads
  • Pinable tooltip
  • Fully separate configuration out from other areas of application, particularly UI
  • Restructure and simplify code behind UI components
  • Add further documentation
  • Adapt to mobile browsers
  • Preload icons (geofilter mode icons, for instance)