Time Series - Breakout detection


Breakout occurs in time series data and have two characteristics:

  • A Mean shift: A sudden jump in the time series corresponds to a mean shift. A sudden jump in CPU utilization from 40% to 60% would exemplify a mean shift.
  • A Ramp up: A gradual increase in the value of the metric from one steady state to another constitutes a ramp up. A gradual increase in CPU utilization from 40% to 60% would exemplify a ramp up.

Time series often contain more than one breakout.

Breakouts detection must be robust, from a statistical standpoint, in the presence of anomalies.


Breakout detection can be used to detect

  • change in user engagement (such as during popular live events such as the Oscars, Super Bowl and World Cup.)
  • hardware issues (breakouts in time series data of system metrics)
  • in user engagement post an A/B test


  • The two red vertical lines denote the locations of the breakouts detected
  • we can see that the detection is robust to anomalies (the peaks)

Twitter R Package

The underlying algorithm of the R package– referred to as E-Divisive with Medians (EDM) – employs energy statistics to detect divergence in mean. Note that EDM can also be used detect change in distribution in a given time series.

See BreakoutDetection on Github

Documentation / Reference

Powered by ComboStrap