Sample Notes: Time Series
O Level and IGCSE Statistics – Detailed Notes: 10. Time Series
10.1 Understanding of Trend
Definition of Trend
- A trend is the general direction in which data points are moving over time.
- It reveals the long-term increase, decrease, or stability in the data.
- Identified by plotting data on a time graph or line chart.
Types of Trends
- Upward Trend: Values are generally increasing over time.
- Downward Trend: Values are generally decreasing over time.
- No Trend: Values fluctuate but show no consistent direction.
- Cyclical/Seasonal Pattern with Trend: Regular fluctuation in data, but with a visible long-term movement.
Example:
Year | Sales (Units) |
---|---|
2018 | 120 |
2019 | 125 |
2020 | 130 |
2021 | 138 |
2022 | 145 |
- Here, there is an upward trend in sales.
10.2 Understanding of Seasonal Variation
Definition
- Seasonal variation refers to periodic fluctuations in data that recur at regular intervals due to seasonal effects (e.g. months, quarters).
- These variations are superimposed on the trend.
Sources of Seasonal Variation
- Weather: e.g. higher ice cream sales in summer.
- Festivals/Events: e.g. higher retail sales in December.
- School terms/Academic calendar
- Economic cycles
Graphical Representation
- On a time series graph, seasonal variations appear as regular spikes and drops around the trend line.
Moving Averages
Purpose
- To smooth out short-term fluctuations in data to better observe the underlying trend.
Simple Moving Average (SMA)
- Calculated by taking the average of a fixed number of consecutive values in the series.
Example: 3-point moving average
Time | Data | 3-Point Moving Average |
---|---|---|
Q1 | 24 | |
Q2 | 27 | |
Q3 | 30 | (24+27+30)/3 = 27.0 |
Q4 | 33 | (27+30+33)/3 = 30.0 |
Q5 | 36 | (30+33+36)/3 = 33.0 |
Centered Moving Average
- Used when the number of periods in the moving average is even.
- Averages two consecutive moving averages to center them.
Example of Centering:
Suppose the 4-point moving averages are:
- (Q1 to Q4) = 28
- (Q2 to Q5) = 30
Then centered moving average at Q2.5 = (28 + 30) / 2 = 29
Seasonal Variation Analysis
Calculating Mean Seasonal Variation
- Determine trend values using moving averages.
- Subtract trend values from actual data to get seasonal components.
- Seasonal Variation = Actual Data – Trend
- Group seasonal variations by period (e.g. Q1, Q2…)
- Calculate the average for each period over multiple years.
Example:
If Q1 data over 4 years:
Actual: 110, 115, 120, 125
Trend values: 108, 112, 118, 123
Seasonal components: 2, 3, 2, 2 → Mean seasonal variation for Q1 = (2+3+2+2)/4 = 2.25
Using Trend and Seasonal Component in Prediction
Forecasting Future Values
- Predicted value = Trend value + Seasonal variation
Steps to Predict:
- Calculate the trend (using moving average or regression).
- Identify seasonal variation for the desired period.
- Add the seasonal component to the trend forecast.
Example:
If the forecasted trend for Q2 next year = 160
And mean seasonal variation for Q2 = +5
Then forecasted value = 165
Time Series Graphs
Components
- X-axis: Time (e.g., months, quarters, years)
- Y-axis: Measured variable (e.g., sales, temperature)
- Plotted points show the actual data
- Trend line can be drawn manually or calculated via moving averages
- Seasonal fluctuations are seen as regular up/down deviations from the trend
Benefits of Time Series Analysis
- Identifies long-term direction of data.
- Highlights repeating patterns in the short-term.
- Allows business forecasting (sales, production, inventory).
- Helps in resource planning and strategic decisions.
Common Mistakes
- Misinterpreting random fluctuations as trends.
- Ignoring the need for centering in even-numbered moving averages.
- Applying wrong seasonal values to wrong quarters.
- Using outdated seasonal patterns for current forecasts.
Exam-Ready Pointers
- Always state whether a trend is upward, downward or none.
- If asked to calculate moving averages: show clear working.
- Use correct units and time intervals.
- For seasonal variation: always mention which quarter/month you’re referring to.
- When drawing time series graphs: label axes clearly and show trend line distinctly.
Key Definitions Recap
Term | Definition |
---|---|
Trend | Long-term direction in data |
Seasonal Variation | Repeating pattern due to time of year |
Moving Average | A method to smooth out short-term variations |
Centered Moving Average | Averaging moving averages to place value between two time points |
Forecast | Prediction using trend and seasonal data |