Gourd

Analytics dashboard framework written in Golang and ChartJS

Problem statement

People need to assemble meaningful visualizations of large, realtime datasets. This problem is partially addressed by large analytics packages, but few of these provide real-time data displays, limiting the ability to proactively respond to new information. Gourd provides a fast, scalable, modern solution.

Storyboard

1. Open page, view a list of streaming data source items (clickable)
2. Select the data sources to stream
3. Selected data charts immediately appear and start real-time streaming
4. Settings:
   • scale/zoom (x and y) (FE-only?)
   • display rate (throttle, FE-only?)
   • chart type: line, bar, area, scatter(?), Front End
   • data selections
5. Time-series data, but could be collapsed to show current value only

Back end TODO

• Buffer for data channel (last 10 minutes or N data points?)

• All graph data from server onload (labels, axes, x & y)

• Multiple simultaneous graphs

• Example data source: Twitter stream (golang twitterstream client)

• performance testing:

  • # points on graph
  • update rate

• heroku deployment

Front end TODO

• Continuous websocket read/timed ui update on client side throttle)
• constant y axis scale
• Pause/resume
• Different graph types
• Gauges?
• fix animation

Features

• pluggable, concurrent data loggers (define API or create admin panel)
• Analytics: min/max
• Analytics: mean

Stack

• Revel framework
• AngularJS
• ChartJS