VCE · General Maths · Data Analysis & Distributionslive from the app
Time Series and Smoothing
A time series plot stacks three things: a long-term trend, a seasonal swing that repeats on the calendar, and short irregular wobble. A centred moving mean strips the wobble, and it cancels the seasonal swing only when its window equals the seasonal period, so the trend stands alone. Two features smoothing cannot erase, and that must be read qualitatively, are a possible outlier from a one-off event and a structural change, where the whole series jumps to a new level and stays there.
This is the real knowscape from knowhere, not a picture of one. Drag it. Watch what actually changes.
general maths · data analysis & distributions · time series and smoothingdrag it · it is yours
the one idea
why this one carries the topic.
Every technique replaces a messy pile of numbers with a shape you can trust — a centre, a spread, a line. You lose the points on purpose: the summary predicts what raw data can't. The topic is deciding which summary the data has earned.
A 48-period series with a season of 6 is plotted with its trend, one seasonal swing and the irregular gap labelled on the plot. A moving-mean window of 6 or 12 cancels the season and leaves the trend; 9 lets the season leak through. Toggles add a labelled outlier, a dashed structural break, or divide the data by its seasonal indices.
what examiners catch — Students use the wrong number of terms in a moving mean, forget to centre an even window, or call a one-off outlier seasonality. Marks come from naming the feature with the right word (trend, seasonality, irregular, outlier, structural change) and matching the smoothing window to the seasonal period.
what you leave with
five things, not forty.
Trend is the long-term direction of the series: read it from the smoothed line over many periods, never from one jump.
Seasonality is a systematic, calendar-related swing: here the same shape repeats every 6 periods, peaking in seasons 1 and 2.
Irregular fluctuations are the short, unsystematic wobbles left between the raw points and the smoothed line.
A possible outlier is a single point far from the pattern, usually traced to a one-off real-world event rather than a season.
Structural change shows as a discontinuity: the level jumps and stays there, so one trend line no longer fits the series.
A moving mean cancels seasonality only when its window equals the seasonal period, and an even window must be centred.
what's underneath
nothing here is a standalone fact.
knowhere maps every concept to what it rests on and what rests on it — 3 underneath this one, 0 built on top. Each one says why, in a sentence, not as an arrow on a diagram.
this concepttime series and smoothingA 48-period series with a season of 6 is plotted with its trend, one seasonal swing and the irregular gap labelled on the plot.
sits under itmeasures of central tendencybecause smoothing IS moving central tendency
sits under itdescribing a distributionbecause Spotting an outlier and describing shape carries over to reading a time series plot.
sits under itleast squares regressionbecause The trend of deseasonalised data is usually fitted with a least squares line.
the rest of data analysis & distributions
9 more, same treatment.
Each one is its own knowscape in the app — built for how a particular student takes things in, not one explanation handed to everybody.
correlation and causationlive →data transformationin the appdescribing a distributionin the appleast squares regressionin the appresiduals and r-squaredin the appthe 68-95-99.7 rulein the appmeasures of central tendencyin the appmeasures of spreadlive →distribution shapes and skewnessin the app
this is one of 865
they don't read it. they run it.
Every HSC and VCE concept in knowhere is built like this one — mapped to your curriculum, never capped by it. Free for a week. No lock-in, cancel inside the week and pay nothing.