A graphical display is a visual presentation of time series observations. Instead of studying only a table of numbers, the analyst plots the data so that pattern, change and unusual behaviour can be seen at a glance.
A graphical display is a visual presentation of time series observations. Instead of studying only a table of numbers, the analyst plots the data so that pattern, change and unusual behaviour can be seen at a glance.
A line chart of monthly electricity demand for two years shows whether demand is rising, whether summer months are regularly higher, and whether any month is unusually high or low.
Graphs are a first step in analysis. A model chosen without looking at the series may ignore trend, seasonality or a sudden change in level. Visual inspection does not replace later numerical checks, but it guides which checks and methods are reasonable.
In simple terms: look at the series before you try to forecast it.
| Feature | What the graph may show | Simple interpretation |
|---|---|---|
| Trend | A long-term rise, fall or stable level | Sales climbing slowly over several years |
| Seasonality | A repeating shape at a known interval | Ice-cream sales peaking every summer |
| Variability | Whether swings become larger or stay similar | Later years showing bigger ups and downs |
| Unusual observations | A spike, drop or isolated odd point | One month of very high sales after a festival |
| Change in behaviour | A break in the earlier pattern | Demand shifting after a new product launch |
Histograms and box plots can summarise the size of values, but they do not keep time order. For time series work, a chronological plot remains the first and most important display.
A shop plots 24 months of packet sales, sees a gentle rise and a dip every February, then decides that both trend and a yearly seasonal pattern should be considered before forecasting.
Always inspect a time series graph before selecting a forecasting method.