A moving average replaces each value with the average of the values in a window around or before it.
It smooths out short-term noise so that the trend is easier to see.
For example, a 7 day moving average of daily sales removes the weekday pattern and shows the underlying direction.
A wider window smooths more but reacts more slowly to genuine changes.
A trailing average uses only past values, so it can be used for forecasting, while a centred one cannot.
Moving averages are the simplest tool for seeing a series' direction through its day to day noise.
They appear everywhere, from sales dashboards to public health reports.
Each point is replaced by the average of a window of neighbouring points.
A trailing window uses the current value and the ones before it, such as the last 7 days.
A centred window uses values on both sides, which lines the average up with the middle of the window.
In pandas, the rolling method computes moving averages, such as a 7 day mean of daily sales.
A window equal to the seasonal period, such as 7 days for a weekly pattern, removes that pattern completely.
Trailing averages lag behind real changes by about half the window length.
A sudden drop therefore shows up in a 28 day average only gradually, over weeks.
Weighted moving averages give recent values more weight, reducing that lag.
The first values of a series have no full window, so they are blank or based on fewer points.
Moving medians resist outliers better than moving averages, since one extreme day barely moves them.
A trailing moving average can serve as a simple forecast, using the latest average as the next value.
In practice, 7 day averages are the standard way to report daily figures with a weekly rhythm.
The usual mistake is a centred average on a dashboard, which needs future values, so the latest days are blank or quietly wrong.
One question on this card. Every option explains itself once you pick, right or wrong.
13 concepts in this module, then a graded test with every answer explained.