Lesson 2 of 7 · 14 min
Trend, seasonality and irregular fluctuations
Describe a time series plot by its trend, its seasonality and its irregular fluctuations, each in the context of the data.
What you will be able to do
- Identify the trend of a time series as increasing, decreasing or no trend.
- Identify seasonality as a systematic, calendar-related pattern and name its peaks and troughs.
- Identify irregular fluctuations and outliers, and describe a plot using all three features in context.
Look at any time series plot long enough and you can see up to three things happening at once. Business at the holiday park is slowly growing. It also rises and falls in the same way every year. And on top of that, each quarter is a little higher or lower than the pattern would suggest, for reasons nobody planned.
Those are the three features you describe on every time series plot.
| Feature | What it is | What to look for |
|---|---|---|
| Trend | The long-term direction of the series. | Is the series generally increasing, decreasing, or showing no trend (flat) over the whole time span? Ignore the ups and downs and look at the overall level. |
| Seasonality | Systematic, calendar-related movements that repeat every cycle (year, week, day). | Peaks and troughs at the same season every cycle — every summer, every December, every Saturday. Name the peak and trough seasons. |
| Irregular fluctuations | Unsystematic, short-term movements with no pattern. | Bumps that don't repeat and can't be predicted. Are they small or large? Is there an outlier — one value far from the pattern, often with a one-off cause? |
Three years of monthly rainfall (data are illustrative). Where are the peaks? Is each year wetter than the last?
Which description best fits the monthly rainfall plot?
Key idea
Describing a time series plot
Comment on all three features, each in context:
- Trend: "There is an increasing / decreasing / no clear trend in [variable] over [time span]."
- Seasonality: "There is seasonality, with peaks in [season] and troughs in [season] each [cycle]" — or "There is no seasonality."
- Irregular fluctuations: "The irregular fluctuations are small / large", and name any outlier with a possible cause.
Worked example
Describe the features of a seasonal time series
The time series plot shows the quarterly guest nights (thousands) at a Sunshine Coast holiday park from Q1 2022 () to Q4 2024 (). Describe the features of the time series.
Quarterly guest nights (thousands), is Q1 2022.
Step 1: Trend
Compare the same quarter across the years: Q1 goes 38, 41, 44 and Q2 goes 21, 23, 24. Every quarter is higher each year, so there is an increasing trend in guest nights.
Comparing the same season each year removes the seasonal swing from the comparison.
Step 2: Seasonality
The same pattern repeats every year: a peak in Q1 (January–March, the summer school holidays), a trough in Q2, then a rise through Q3 and Q4. So there is seasonality with a yearly cycle.
Step 3: Irregular fluctuations
The pattern is almost identical each year, so the irregular fluctuations are small. There are no outliers.
Try it yourself first, then check the working one step at a time.
Answer
There is an increasing trend in guest nights from 2022 to 2024. There is seasonality, with a peak every Q1 (summer holidays) and a trough every Q2. The irregular fluctuations are small and there are no outliers.
Method
Trend: compare the overall level, or the same season, across cycles.
Seasonality: find the season that peaks and the season that dips in every cycle.
Irregular: how closely does each cycle follow the pattern? Name any outlier.
Watch out for
Saying "the trend goes up and down". The up-and-down movement is the seasonality; the trend is the long-term direction.
Describing the features without the context (guest nights, quarters, summer holidays).
Mean daily water use per person in a south-east Queensland town (litres), 2015–2024. Data are illustrative.
Describe the water-use time series plot.
Note
Seasonality needs more than one value per cycle
Seasonality is a pattern within a cycle. Quarterly or monthly data can show a yearly pattern; daily data can show a weekly pattern (a café's busy weekends); hourly data can show a daily pattern (electricity demand peaking at 6 pm). Annual data has one value per year, so it can't show a yearly seasonal pattern at all.
An outlier is an irregular fluctuation so large that it stands out from everything else, usually with a one-off cause: a cyclone, a flood, a strike, a pandemic. In the agricultural-show data from the last lesson, 2020 attendance fell to 18.2 thousand when the show ran with heavy restrictions. That is not part of the trend and won't repeat each year; it is an outlier.
A cyclone closed the holiday park for two weeks in March 2023, so Q1 2023 was lower than expected. How would you classify this effect?
Common mistake
Calling every rise and fall seasonality
Seasonality must be systematic and calendar-related: the same seasons are high and the same seasons are low in every cycle. Random rises and falls are irregular fluctuations. And a seasonal rise is not a trend — a series can rise every summer and still have no trend at all, as the rainfall does.
Exam tip
Three features, three marks
A "describe the time series" question usually has a mark for each feature. Write one clear sentence per feature in context, and name the peak and trough seasons. If a feature is absent, say so ("there is no seasonality") — an omitted feature earns nothing.
Summary
Trend: the long-term direction — increasing, decreasing or no trend.
Seasonality: a systematic, calendar-related pattern that repeats each cycle; name the peak and trough seasons.
Irregular fluctuations: unsystematic, short-term movements; an outlier is an extreme one, usually with a one-off cause.
Annual data can't show a yearly seasonal pattern; you need several values per cycle.
Describe all three features, each in context, in a full sentence.
Finished the lesson?
Try the 3 quick checks above and the lesson completes itself, or mark it complete now.