Time series analysis

A trend in a time series is upward or downward, and a linear trend can be quantified by least-squares fit to a straight line. However, the trend need not be linear; some trends cannot be linear. A non-linear example is a drop in birth rate. A straight line sloping downward would reach zero, but the birth rate must level off (unless we are studying some endangered species headed for extinction).

Recognizing periodicity in a time series can be valuable. You may be able to anticipate several years of drought or of excessive rain. We seem to be in a period of global warming on a time scale that includes the retreat of continental glaciers, but man is accelerating the effects by generating vast amount of carbon dioxide and by releasing chemicals that damage the ozone layers.
See how moving averages can clarify relationships.
Graph with a moving average.

Fluctuations are changes with no firm pattern. The characteristics to look for are the frequency and the magnitude of the changes. Some events such as firing of nerve cells or earth movements due to earthquakes may appear to be random fluctuations while in fact the intervals and intensities convey information about process conditions that trigger them.

 More about time series