NCEA Level 1Time seriesAS 91946

Time series graphs: trend, seasonal pattern and unusual points

What this lesson covers

  • A time series graph shows how one variable changes over time, and its features have names.
  • Trend, seasonal pattern and unusual spikes are read separately and then described together.
Key idea

A time series graph

  • One variable, measured at regular times.
  • Time always runs across.
  • Often lost marks in the exam.
Shows
Acrosstime, in quarters1 to 24, six years
Upvisitors, thousands9 to 24
Q1January to Marchsummer
Q3July to Septemberwinter
Worked example

Reading values first

  • Read a value: go up from the time, across to the scale.
  • Include the units: thousands of visitors.
  • A change is one value minus another.
ValueWhen
Quarter 116 thousandfirst summer
Quarter 310 thousandfirst winter
Change16 − 10 = 6 thousandsummer to winter
Worked example

Six years of visitors

  • A zigzag that climbs.
  • Each zig is one year.
  • Three features to read, separately.
Six years of visitors
Worked example

Feature 1: the trend

  • Ignore the zigzag. Where is it heading overall?
  • Compare like with like: summer to summer.
  • Up 8 thousand in 5 years: 1.6 a year.
Feature 1: the trend
Worked example

The trend, with numbers

  • 16 thousand in the first summer.
  • 24 thousand in the last.
  • A rate makes the trend precise.
Visitors
Summer, year 116 thousand
Summer, year 624 thousand
Change8 thousand in 5 years1.6 a year
Key idea

Feature 2: the seasonal pattern

  • The same shape repeats every four quarters.
  • Highest in quarter 1, lowest in quarter 3.
  • Give the size of the swing: about 6 thousand.
When
Peakquarter 1 every yearsummer
Troughquarter 3 every yearwinter
Swingabout 6 thousandevery year
Worked example

Feature 3: an unusual point

  • Quarter 15 drops to 9 thousand.
  • The pattern says about 15 thousand.
  • Something happened: suggest what.
Feature 3: an unusual point
Key idea

Explaining it from context

  • A road closure, an exhibit shut, bad weather.
  • Say "possibly": the graph cannot tell you.
  • Do not let one dip change the trend.
Worked example

All three, in context

  • Each feature gets a sentence.
  • Each sentence gets a number.
  • Each number gets its units.
FeatureDescription
Trendrising, about 1.6 thousand a year
Seasonalpeaks in summer, lowest in winter
Sizea swing of about 6 thousand
Unusualquarter 15: 9 thousand, about 6 low
Worked example

A careful prediction

  • Start from the latest same season.
  • Add a year of trend: 24 + 1.6.
  • About 25 to 26 thousand next summer.
A careful prediction
Key idea

Why only "about"

  • A prediction assumes the pattern continues.
  • Quarter 15 shows it sometimes does not.
  • Predict one season ahead, not ten years.
Last summer24 thousand
Trend+1.6 thousand a year
Next summer24 + 1.6 = 25.6about 25 to 26 thousand
Key idea

How the grades step up

  • Achieved: one supported feature, in context.
  • Merit: features with numbers from the graph.
  • Excellence: reasons, prediction and caution.
Summary

What to take away

  • Trend, season, unusual: each with a number.
Time seriesone variable over time
Trendlong-term direction, with a rate
Seasonalwhen it peaks, when it dips, how much
Unusualname it, give a possible reason
Separatedo not let a dip hide the trend
Predicttrend plus season, cautiously
Numbersfrom the graph, in context
Your turn

Your turn: four questions

  • Sketch the series if it helps.
  • Pause here and answer each one.
  • Compare like quarters with like.
Question
DataIce block sales ($1000s), Q1 to Q4. Y1: 18 9 5 12. Y2: 20 11 7 14. Y3: 22 13 9 16.
1In which quarter do sales peak each year?
2Describe the trend, with numbers.
3Predict sales for Q1 of year 4.
4Give one reason the prediction could be wrong.
Show answer

1. Quarter 1, January to March, the summer quarter. 2. Rising: every quarter is $2000 higher than the same quarter the year before, and yearly totals rise from $44 000 to $52 000 to $60 000. 3. 22 + 2 = 24, so about $24 000. 4. For example: a cool, wet summer, a new competitor, or a price change; the prediction assumes the pattern continues.

Worked example

The answers

  • Quarter 1, the summer quarter.
  • Up $2000 a quarter each year; totals 44, 52, 60.
  • 22 + 2: about $24 000.
  • e.g. a wet summer: it assumes the pattern holds.

Is your child on track?

Take the free placement test