Time series is essentially data that is recorded in time order. In nearly all cases, the data includes a timestamp with one or more data points being collected at that time. Time series data is commonly seen in fields that analyze trends over time.

Characteristics

  • Trends: Drifting up or down
  • Seasonality: Regularly occurring
  • Cyclic: Exhibits a rise and fall at unknown intervals
  • Stationary: Mean & std dev is constant
    • Trend-stationary is also possible

Tests

  • Engle Granger Method
    • Useful for determining if there’s a correlation between two processes
  • KPSS test
    • Hypothesis test to determine if the process is stationary or trend-stationary