Guide to statistics

Time series data is a collection of observations obtained at regular intervals over time. It allows changes to be identified and can show the impact of cyclical, seasonal or irregular events on the data being measured. Time series data can be classified into two different types: stock and flow. A stock series measures certain attributes at a point in time such as the Australian Bureau of Statistics (ABS) Australian Demographic Statistics Cat. No. 3101.0, which measures the population at a point in time. A flow series measures activity over a given period accumulated during the reference period, such as the ABS Retail Trade, Australia Cat. No. 8501.0, which measures monthly retail turnover.

Data collected is generally analysed in three ways, original, seasonally adjusted or trend. Original data shows the actual movements in the data over time and it may include movements due to cyclical, seasonal or irregular events. Seasonally adjusted data involves the removal of cyclical and seasonal effects from the original data. For example, employment data is often seasonally adjusted, to remove the impact of periods of known peak employment such as Christmas when a large number of casual workers are temporarily employed.

Trend data has further adjustments made to seasonally adjusted data to remove irregular effects. Irregular effects can include major short-term events such as flood or the introduction of a new tax, industrial disputes as well as the impact of the occasional large project. The irregular fluctuations can also be due to sampling and non-sampling errors. Trend data is the long-term movement in a time series without calendar-related and irregular effects, and is reflecting the underlying level. It also captures various medium-term business cycles.

Since the Northern Territory (NT) has a small population, relatively small economy and - many remote areas, there are issues with data collected for the NT. Small sample sizes for can result in data being skewed as it may not be truly representative of the population and it can lead to greater volatility. It may be difficult to collect data due to remoteness. NT’s relatively small economy and population size can lead to data being made confidential. This causes volatility in the data collected for the NT. Trend data is often used for analysis as it smooths seasonal and one off variations. For the calculation of annual change, which is comparing the change from one point in time with the same time in the previous year, trend data is used.

As the methodology the ABS uses to calculate the trend data (that is, the last six months of published data) is not actuals but forecasts, it can lead to substantial revisions of the last six months’ data especially for smaller jurisdictions. Original data is used to measure the level of activity over the period of a year for the NT. Original data is not suitable to compare one period to the next as it is dominated by seasonal effects and irregular influences. Year-on-year change is calculated by comparing the data for the most recent 12 months or four quarters (depending on the frequency of data collection) against the previous 12 months or four quarters. This is done by using the moving annual total or moving annual average for the data for each of the years, which involves adding up the data for the year or averaging the data for the year.

Chain volume or inflation-adjusted measures provide estimates of real changes by factoring general changes in prices from year to year. This allows comparison of expenditure or production levels that are free of the direct effects of price change. Chain volume measures are derived by linking together movements in volumes, calculated using the average prices of the previous financial year and applying the compounded movements to the current price estimates of the reference year.

For further information on statistical measures and terminology please go to the Australian Bureau of Statistics website.