Skip to content

New Data Quality Algorithm

Gabriel Jacoby-Cooper edited this page Oct 17, 2022 · 9 revisions

What’s New

  • Data quality will be determined by time instead of by source (currently, user data is green and bus GPS data is red)
  • Data source is not a factor at all—that is, user data will be treated no differently from bus GPS data
  • Quality is stored in a float from 0.0 to 1.0

Clients

  • Highest quality is 1.0 (should be displayed in green 🟩)
  • Low quality is 0.4 after 3 minutes (should be displayed red 🟥)
  • After 5 minutes, data are deemed lowest quality (0.0)
  • Quality formula: [Quality] = 1 - ([Seconds Elapsed] / 5 * 60)
  • Clients can choose how to display color: either as gradient as quality decreases or as strictly red or green

Server

  • When submitting a location to clients, don’t average the coordinate values over the data points; instead, just return the newest datum (with the highest quality)
  • Quality will be updated every 10 seconds
  • See this issue